100 Free Cloud+ CV0-004 Practice Questions and Answers
100 questions for the Cloud+ CV0-004 exam, each answer explained. Watch it here, try the questions yourself, or read the transcript.
Questions in this video
Tap an answer to see why it is right or wrong. Tap it again to reset.
Question 1 Play from 00:00
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
Each bookstore would pay for its own dedicated setup, which drives the price up.
Right, B is the answer.
Why B is right
When a handful of organizations in the same line of business have similar needs, they can split the bill for one shared environment that only they use. That gives them savings from pooling plus more control than a general-purpose public offering.
Not this one. Try another answer.
Why C is wrong
It is open to everyone and gives this group no special shared arrangement among themselves.
Not this one. Try another answer.
Why D is wrong
It mixes on-premises and cloud resources, which adds complexity the group does not need.
Question 2 Play from 01:29
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
It follows a calendar, so it cannot react to an unexpected jump in usage.
Right, B is the answer.
Why B is right
Scaling based on load watches live utilization and adds capacity the moment a threshold is crossed. That is exactly what you want when usage rises past a set percentage, with no person or calendar involved.
Not this one. Try another answer.
Why C is wrong
It depends on someone noticing and acting, which is slow.
Not this one. Try another answer.
Why D is wrong
It looks at past patterns to predict demand rather than reacting to current utilization.
Question 3 Play from 02:51
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
It is a point-in-time copy used for recovery, not a way to run workloads.
Not this one. Try another answer.
Why B is wrong
You still have to build images, store them in a registry, and usually run an orchestrator.
Not this one. Try another answer.
Why C is wrong
You must set up nodes and orchestration before anything runs.
Right, D is the answer.
Why D is right
With this model you hand over your code and the provider takes care of servers, scaling, and capacity for you. There is almost nothing to configure, so it is the quickest path from idea to running service.
Question 4 Play from 04:09
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
It interprets text that is already in digital form and does not read paper.
Right, B is the answer.
Why B is right
This technology, often called OCR, looks at the image of paper and converts the characters into text a computer can use. The other choices either only capture a picture or work on text that is already digital.
Not this one. Try another answer.
Why C is wrong
It judges the mood or opinion in text rather than reading it from paper.
Not this one. Try another answer.
Why D is wrong
It produces an image of the page but does not pull out the words.
Question 5 Play from 05:32
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
It examines application-level content, which adds processing time.
Not this one. Try another answer.
Why B is wrong
It is an isolated network space, not a service for distributing and publishing traffic.
Right, C is the answer.
Why C is right
A load balancer working at the transport layer passes traffic along without inspecting the content of each request. Because of that light touch it adds very little delay and can handle huge connection volumes.
Not this one. Try another answer.
Why D is wrong
It makes decisions on application data, so it adds more delay than a transport-layer option.
Question 6 Play from 06:57
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
It is a test-stage pre-release, not a production update.
Not this one. Try another answer.
Why B is wrong
That would show up as a change in the first number, such as 3.2 to 4.0.
Not this one. Try another answer.
Why C is wrong
It is an early, unfinished build and not a production release.
Right, D is the answer.
Why D is right
Under the common numbering scheme, bumping the second number signals new capabilities that still work with existing users. A jump in the first number would mark a major change instead, and the pre-release labels apply to unfinished software.
Question 7 Play from 08:17
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
It disappears when the instance stops, so it cannot safely hold database data.
Right, B is the answer.
Why B is right
Databases do a lot of small random reads and writes and need quick, direct access to disk-like volumes. Block volumes attach to a server and deliver that low delay and steady throughput.
Not this one. Try another answer.
Why C is wrong
It suits large unstructured files and has higher latency, which hurts database work.
Not this one. Try another answer.
Why D is wrong
It is built for shared access to files, not for the random access patterns of a database.
Question 8 Play from 09:36
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
Each machine still needs its own OS installed and maintained.
Not this one. Try another answer.
Why B is wrong
Every virtual machine carries its own guest OS that someone has to look after.
Right, C is the answer.
Why C is right
Containers share the host's operating system kernel, so many workloads can run without each one needing its own full OS. That saves resources and leaves far fewer systems to patch and maintain.
Not this one. Try another answer.
Why D is wrong
They add more operating system instances to manage rather than fewer.
Question 9 Play from 10:56
Tap an answer. The reason for that answer shows here.
Right, A is the answer.
Why A is right
This standard was created by the card brands and sets rules for anyone handling cardholder data, including protecting it with encryption while stored and while moving across networks.
Not this one. Try another answer.
Why B is wrong
It is an audit report on service provider controls, not a card data rulebook.
Not this one. Try another answer.
Why C is wrong
It is a broad European privacy law and is not specific to card numbers.
Not this one. Try another answer.
Why D is wrong
It publishes general cloud security guidance and does not set card data rules.
Question 10 Play from 12:24
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
It concerns who holds rights and responsibility for the data, not where it sits.
Right, B is the answer.
Why B is right
This idea says data falls under the laws of the place where it is stored or collected, so it often has to stay in that country or region. It is a legal matter tied to location.
Not this one. Try another answer.
Why C is wrong
It is about how long data is kept before being removed.
Not this one. Try another answer.
Why D is wrong
It sorts data by sensitivity level, not by geography.
Show all 100 questionsShow fewer questions
Question 11 Play from 13:43
Tap an answer. The reason for that answer shows here.
Right, A is one of the answers.
Why A is right
With a managed database, the provider runs the underlying platform, while the customer still decides who can see which data and how sensitive records are protected. Access rules on tables and encryption of individual rows both fall on the customer side.
Not this one. Try another answer.
Why B is wrong
The provider maintains the operating system under the managed service.
Not this one. Try another answer.
Why C is wrong
Supplying and scaling physical capacity is the provider's job.
Right, D is one of the answers.
Why D is right
With a managed database, the provider runs the underlying platform, while the customer still decides who can see which data and how sensitive records are protected. Access rules on tables and encryption of individual rows both fall on the customer side.
Not this one. Try another answer.
Why E is wrong
The provider applies engine updates as part of the managed offering.
Not this one. Try another answer.
Why F is wrong
The provider sets up and runs the cluster behind the service.
Question 12 Play from 15:23
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
It joins work from different lines together but does not create the separate tracks.
Not this one. Try another answer.
Why B is wrong
It only sends commits to a remote repository.
Right, C is the answer.
Why C is right
Creating a separate line of development lets each customer's variation evolve on its own while still starting from the shared product code. Changes can later be brought back together when appropriate.
Not this one. Try another answer.
Why D is wrong
It saves a set of changes but does not separate work per customer.
Question 13 Play from 16:42
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
Its hierarchical design does not scale or save money as well at data lake size.
Not this one. Try another answer.
Why B is wrong
It is temporary and disappears, so it cannot hold data that must be kept.
Not this one. Try another answer.
Why C is wrong
It fits databases and servers needing fixed volumes, not massive unstructured collections.
Right, D is the answer.
Why D is right
Data lakes hold enormous amounts of mixed, mostly unstructured data. Object storage scales almost without limit, costs less per unit, is durable, and plugs easily into analytics tools.
Question 14 Play from 18:00
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
It gives vulnerabilities an identifier and a catalog entry but no severity score.
Not this one. Try another answer.
Why B is wrong
It scores types of software weaknesses, not a vulnerability in a running system.
Right, C is the answer.
Why C is right
This system turns the traits of a known vulnerability into a number, which helps teams decide what to fix first. The other acronyms name or categorize problems rather than rating how bad a specific flaw is.
Not this one. Try another answer.
Why D is wrong
It is a list of weakness categories and has no severity calculation.
Question 15 Play from 19:28
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
Processing capacity is a sizing detail that follows from the security and legal needs.
Right, B is one of the answers.
Why B is right
Health organizations handle private patient records and must follow strict regulations. That makes protecting the data and meeting legal obligations the top concerns, ahead of the technical sizing choices that follow from them.
Not this one. Try another answer.
Why C is wrong
Budget matters, but it ranks below protecting patient data and meeting regulations.
Right, D is one of the answers.
Why D is right
Health organizations handle private patient records and must follow strict regulations. That makes protecting the data and meeting legal obligations the top concerns, ahead of the technical sizing choices that follow from them.
Not this one. Try another answer.
Why E is wrong
Capacity is planned after the security and compliance requirements are set.
Not this one. Try another answer.
Why F is wrong
Uptime is valuable but is not the leading concern compared with protection and regulation.
Question 16 Play from 21:10
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
It delivers finished software and gives no way to deploy their own code.
Right, B is the answer.
Why B is right
Renting raw virtual servers by the hour lets the team spin up a large fleet for a short test and release it afterward. They keep full control of what runs on those machines and pay no long-term commitment.
Not this one. Try another answer.
Why C is wrong
It only provides managed databases, not general servers.
Not this one. Try another answer.
Why D is wrong
It hides the servers, which limits the control needed to build a custom large cluster.
Question 17 Play from 22:30
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
It belongs to Fedora and newer Red Hat style systems.
Right, B is the answer.
Why B is right
Ubuntu comes from the Debian family, and that family relies on this tool to install, update, and remove software.
Not this one. Try another answer.
Why C is wrong
It is used on older Red Hat style distributions.
Not this one. Try another answer.
Why D is wrong
It is the package tool for Arch Linux.
Question 18 Play from 23:37
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
Build agents run pipeline jobs and do not define or enforce how infrastructure is built.
Not this one. Try another answer.
Why B is wrong
It gathers and searches logs but does not provision infrastructure.
Right, C is the answer.
Why C is right
Writing the infrastructure once as code and feeding it different variable values means every environment is created from the same blueprint. Only intended settings such as size or count differ, which removes drift between them.
Not this one. Try another answer.
Why D is wrong
It visualizes metrics and has no role in making environments match.
Question 19 Play from 25:21
Tap an answer. The reason for that answer shows here.
Right, A is the answer.
Why A is right
An application-layer balancer can read details such as the hostname and URL path, so it can send each request to the right backend while spreading the work across the pool of machines.
Not this one. Try another answer.
Why B is wrong
It filters malicious traffic but does not distribute requests across servers.
Not this one. Try another answer.
Why C is wrong
It acts for clients going outward, not for servers receiving inbound traffic.
Not this one. Try another answer.
Why D is wrong
It caches content near users and does not route among backend virtual machines.
Question 20 Play from 26:43
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
It stacks multiple layers of protection but is a general strategy, not a per-access verification model.
Right, B is the answer.
Why B is right
This model assumes nothing is safe by default and checks identity, permissions, location, and device status on every access attempt. That lines up with each of the listed conditions.
Not this one. Try another answer.
Why C is wrong
It is about building security into a system from the start, not checking users and devices.
Not this one. Try another answer.
Why D is wrong
It limits permissions to what is needed but says nothing about location or device checks.
Question 21 Play from 28:12
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
It recopies every change since the last full backup, so it grows larger and runs longer than an incremental job.
Right, B is the answer.
Why B is right
The incremental approach copies only what changed since the previous backup of any kind, so every run is tiny and quick. A differential job keeps growing because it recopies everything since the last full one. Full jobs copy the whole data set each time, and tape is a storage medium rather than a backup method.
Not this one. Try another answer.
Why C is wrong
Tape is a storage medium, not a backup type, and it is usually slower.
Not this one. Try another answer.
Why D is wrong
It copies the entire data set on every run, which is the largest and slowest choice.
Question 22 Play from 29:44
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
It records events inside one server, so it cannot show the overall traffic pattern of an attack.
Not this one. Try another answer.
Why B is wrong
These track malware and suspicious behavior on hosts, not floods of network traffic.
Right, C is the answer.
Why C is right
Flow logs record who is talking to whom, on which ports, and how much data moves. A flood of requests from many sources shows up there as a sharp, unusual jump in volume. That makes them the natural place to spot a distributed denial-of-service attack.
Not this one. Try another answer.
Why D is wrong
They record account and API actions, not the packets hitting the workload.
Question 23 Play from 31:23
Tap an answer. The reason for that answer shows here.
Right, A is the answer.
Why A is right
The app is both too slow to scale and costly to keep alive, so it needs deeper changes than a simple move. Refactoring reworks the code and design for scale and easier upkeep, which solves both complaints.
Not this one. Try another answer.
Why B is wrong
It swaps some platform pieces but keeps the same core design, so scaling and upkeep problems can remain.
Not this one. Try another answer.
Why C is wrong
Leaving the app as it is does nothing for demand or maintenance cost.
Not this one. Try another answer.
Why D is wrong
A lift-and-shift copies the app to new infrastructure unchanged, so the same limits and costs come along.
Question 24 Play from 32:53
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
It serves data that is used all the time with fast access.
Right, B is the answer.
Why B is right
The archive tier is built for long-term records that are written once and fetched only when someone asks, at the lowest cost per gigabyte. Hot data is used constantly and warm data now and then, while cold data is infrequent but is not defined as read-only retention.
Not this one. Try another answer.
Why C is wrong
It holds data touched occasionally and is not defined by being read-only.
Not this one. Try another answer.
Why D is wrong
It is for infrequent access, but it is not defined as read-only retention storage.
Question 25 Play from 34:17
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
A fix cannot be chosen sensibly until the problem is understood.
Not this one. Try another answer.
Why B is wrong
Calling others in comes after a first assessment shows it is needed.
Not this one. Try another answer.
Why C is wrong
Tuning monitoring is a follow-up task, not the immediate reaction to an outage.
Right, D is the answer.
Why D is right
The first move on a critical page is to size up the situation: how bad it is, how many users are hit, and what is likely causing it. That assessment is triage, and it guides every step after it, including whether to fix things directly or call for help.
Question 26 Play from 35:44
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
It adjusts resource sizes to save money but does not label costs.
Right, B is the answer.
Why B is right
Labels attached to individual resources, such as project or owner, flow into billing data so costs can be filtered and grouped per resource. Rightsizing and reserved purchasing are ways to save money later, and an invoice only shows totals.
Not this one. Try another answer.
Why C is wrong
An invoice summarizes charges overall without breaking them down by resource.
Not this one. Try another answer.
Why D is wrong
They give a discount for a usage commitment and add no per-resource cost detail.
Question 27 Play from 37:07
Tap an answer. The reason for that answer shows here.
Right, A is the answer.
Why A is right
A transit gateway acts as a central hub that many VPCs and an on-premises link can all plug into, and it passes traffic between them. Peering only joins two VPCs directly, a subnet lives inside one VPC, and a CDN is for caching content near users.
Not this one. Try another answer.
Why B is wrong
It is just a slice of address space inside a single VPC.
Not this one. Try another answer.
Why C is wrong
It speeds up content delivery and does not link private networks.
Not this one. Try another answer.
Why D is wrong
It joins two VPCs directly and does not provide hub routing across many networks or to on-premises.
Question 28 Play from 38:35
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
It is open to anyone, so need-to-know limits cannot be guaranteed.
Not this one. Try another answer.
Why B is wrong
It is shared by several organizations with common interests, which widens access.
Not this one. Try another answer.
Why C is wrong
It blends in public cloud resources, which brings in outside exposure.
Right, D is the answer.
Why D is right
A private cloud serves one organization only, and it can sit inside its own facilities with tightly controlled need-to-know permissions. Sharing with other organizations, opening to the internet, or mixing in public resources all weaken that control.
Question 29 Play from 40:01
Tap an answer. The reason for that answer shows here.
Right, A is the answer.
Why A is right
Reports that look far back in time need the data to still exist, so the length of time backups are kept is the key setting. Everything else tunes how backups are made or stored, not how long they survive.
Not this one. Try another answer.
Why B is wrong
Storing backups elsewhere changes where data sits, not how long it remains available.
Not this one. Try another answer.
Why C is wrong
Full, incremental, or differential affects speed and size, not how long data is preserved.
Not this one. Try another answer.
Why D is wrong
Frequency decides how many restore points exist, not how long they are kept.
Question 30 Play from 41:33
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
It shares the host operating system kernel rather than bringing its own full OS.
Not this one. Try another answer.
Why B is wrong
A cluster is a group of machines working together, not a single isolated compute unit.
Not this one. Try another answer.
Why C is wrong
A stand-alone server is just one physical machine and does not describe a virtualized, isolated unit.
Right, D is the answer.
Why D is right
A virtual machine emulates a full computer on top of physical hardware, with its own virtual network, operating system, processor share, and memory. That gives each workload strong isolation from its neighbors.
Question 31 Play from 43:03
Tap an answer. The reason for that answer shows here.
Right, A is the answer.
Why A is right
A content delivery network keeps copies of the static pages at edge locations spread around the globe, so each visitor is served from a nearby point instead of one distant origin. That shortens the delay. A VPC only isolates a network, a load balancer spreads requests among servers in one place, and an API gateway fronts API calls.
Not this one. Try another answer.
Why B is wrong
It provides a private network boundary and does nothing to bring content closer to visitors.
Not this one. Try another answer.
Why C is wrong
It manages and secures API calls, not delivery of static site files worldwide.
Not this one. Try another answer.
Why D is wrong
It spreads traffic among servers in a region and does not shorten the distance to far-away users.
Question 32 Play from 44:35
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
The team would have to patch, back up, and maintain the server and the database software itself.
Not this one. Try another answer.
Why B is wrong
It is a tool for moving existing data, not a place to run a new application's database.
Right, C is the answer.
Why C is right
With a managed relational service, the provider takes care of patching, backups, failover, and scaling, so the team simply uses the database. Running a database engine on a server leaves all that work with the team, a hybrid layout adds more moving parts, and a migration service only moves data from one place to another.
Not this one. Try another answer.
Why D is wrong
Spanning cloud and local systems adds complexity and more administration.
Question 33 Play from 46:23
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
It makes copies on a schedule, but nothing stops those copies from being altered or removed early.
Not this one. Try another answer.
Why B is wrong
It duplicates data to other places, including any unwanted changes, and enforces no retention.
Not this one. Try another answer.
Why C is wrong
It is about restoring lots of data after a failure, not about protecting records for compliance.
Right, D is the answer.
Why D is right
Object Lock gives write-once, read-many protection on cloud bucket objects, so records cannot be changed or deleted until the retention period, such as five years, ends. Combined with access controls for privacy, that meets every item on the list. The other choices copy data but cannot enforce immutability or a retention period.
Question 34 Play from 48:00
Tap an answer. The reason for that answer shows here.
Right, A is the answer.
Why A is right
The application starts in a local data center and needs to reach outside users, so the move is from on-premises into a cloud. Cloud providers offer wide geographic reach and redundant capacity that suit a highly available service for external audiences.
Not this one. Try another answer.
Why B is wrong
That is the opposite direction and would not help reach outside users.
Not this one. Try another answer.
Why C is wrong
Staying in local data centers gives no easier way to reach external users.
Not this one. Try another answer.
Why D is wrong
That applies when the workload already lives in a cloud, which is not the starting point here.
Question 35 Play from 49:33
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
A set amount that never changes is the opposite of elastic.
Not this one. Try another answer.
Why B is wrong
Encryption protects data confidentiality and has nothing to do with scaling.
Right, C is the answer.
Why C is right
Elasticity is the ability to add resources when load rises and release them when it falls, without manual work. Fixed capacity, encryption, and backup timing are different topics.
Not this one. Try another answer.
Why D is wrong
How often backups run is a data protection setting, not a scaling behavior.
Question 36 Play from 51:02
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
Developers would still have to build and manage the virtual machines and the software stack.
Right, B is the answer.
Why B is right
With a platform service, the provider runs the servers, operating systems, runtimes, and tooling, leaving developers free to write and deploy code. IaaS hands the servers back to manage, FaaS covers only small event-driven functions, and SaaS is finished software that is not meant for custom development.
Not this one. Try another answer.
Why C is wrong
It suits small event-triggered functions and is not a full platform for general applications.
Not this one. Try another answer.
Why D is wrong
It delivers ready-made software to use, not an environment to build custom code.
Question 37 Play from 52:41
Tap an answer. The reason for that answer shows here.
Right, A is the answer.
Why A is right
Serverless runs code only when a request arrives and bills for that execution time, so a rarely used service costs almost nothing while idle, and it still scales and responds quickly. The other choices keep capacity reserved or running all the time and charge for it.
Not this one. Try another answer.
Why B is wrong
Containers need a host and allocated resources that keep costing money between requests.
Not this one. Try another answer.
Why C is wrong
Edge deployments rely on dedicated distributed infrastructure that is paid for continuously.
Not this one. Try another answer.
Why D is wrong
A virtual machine is billed while it runs, whether or not anyone uses it.
Question 38 Play from 54:10
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
A review checks the work for quality, but it does not itself store the change in history.
Not this one. Try another answer.
Why B is wrong
Refactoring restructures existing code and is not how a fix gets recorded.
Not this one. Try another answer.
Why C is wrong
This is not a standard version control step.
Right, D is the answer.
Why D is right
Committing saves the fix into version control together with a message, which adds it to the permanent project history. Reviews, refactors, and requests do not by themselves create that record.
Question 39 Play from 55:29
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
Sharing a cluster with the main site is a local high availability design and not what defines a hot site.
Right, B is the answer.
Why B is right
A hot site mirrors the production environment closely: every server is copied over from the main site and stays up and running, so failover can happen almost instantly. Warm and cold sites hold offline or partly prepared gear instead.
Not this one. Try another answer.
Why C is wrong
Offline hourly copies describe a warm or cold approach.
Not this one. Try another answer.
Why D is wrong
Splitting traffic across both sites is an active-active design, not the defining trait of a hot site.
Question 40 Play from 57:16
Tap an answer. The reason for that answer shows here.
Right, A is the answer.
Why A is right
Spot capacity is spare provider capacity sold at steep discounts, and the provider may take it back, which is fine for brief, low-stakes jobs. Reserved pricing needs a long commitment, pay-as-you-go costs more per hour, and dedicated hosts are the priciest.
Not this one. Try another answer.
Why B is wrong
It rewards a long commitment, which does not fit short-term work.
Not this one. Try another answer.
Why C is wrong
On-demand rates are higher than spare-capacity discounts.
Not this one. Try another answer.
Why D is wrong
A whole physical server is set aside for you, which is the most expensive route.
Question 41 Play from 58:38
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
Passthrough gives a virtual machine direct use of a physical device and does nothing for uptime.
Not this one. Try another answer.
Why B is wrong
A lone container is a single point of failure and cannot span regions by itself.
Right, C is the answer.
Why C is right
Grouping several nodes that sit in different regions lets traffic shift automatically when one of them fails, so the service keeps answering quickly. The other choices either copy a resource, expose a device, or leave a single point of failure.
Not this one. Try another answer.
Why D is wrong
Cloning only duplicates a resource, with no failover or geographic spread.
Question 42 Play from 1:00:00
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
The vendor sets up the service as part of the managed offer.
Right, B is the answer.
Why B is right
With a managed offering, the vendor takes care of the build, the server updates, and the setup. Deciding who may reach the service and what each person may do stays with the customer, so identity and access settings are the item to plan.
Not this one. Try another answer.
Why C is wrong
Building software from source is hidden behind the managed service.
Not this one. Try another answer.
Why D is wrong
The provider keeps the underlying operating system patched.
Question 43 Play from 1:01:26
Tap an answer. The reason for that answer shows here.
Right, A is one of the answers.
Why A is right
Banks handle sensitive money data and answer to strict rules, so protecting that data and proving adherence to the rules come before everything else. Price, speed, and raw capacity still matter, but they follow after those two.
Not this one. Try another answer.
Why B is wrong
Spend matters, yet it ranks below protection and rule-following for a bank.
Not this one. Try another answer.
Why C is wrong
Network design supports the top goals but is not itself the leading demand.
Right, D is one of the answers.
Why D is right
Banks handle sensitive money data and answer to strict rules, so protecting that data and proving adherence to the rules come before everything else. Price, speed, and raw capacity still matter, but they follow after those two.
Not this one. Try another answer.
Why E is wrong
Processing power is a general infrastructure need, not specific to finance.
Not this one. Try another answer.
Why F is wrong
Speed is nice to have but gives way to regulation and data protection.
Not this one. Try another answer.
Why G is wrong
Capacity matters, but the controls around the stored data matter more.
Question 44 Play from 1:03:12
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
It interprets images and video, not the mood of written words.
Right, B is the answer.
Why B is right
This technology reads written language and decides whether the mood is positive, negative, or neutral. That is exactly the job of judging how a message feels, which the other options cannot do.
Not this one. Try another answer.
Why C is wrong
It spots things in pictures rather than reading tone in text.
Not this one. Try another answer.
Why D is wrong
It pulls characters out of images, which is OCR, and says nothing about mood.
Question 45 Play from 1:04:36
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
These are provider facility costs that do not shift with the customer's layout.
Right, B is the answer.
Why B is right
Traffic that crosses region borders is billed as outbound transfer, and a multi-region design generates plenty of it. A single region avoids most of those charges, so networking is where the bill grows fastest.
Not this one. Try another answer.
Why C is wrong
Replicated copies add some storage cost, but transfer is the bigger jump.
Not this one. Try another answer.
Why D is wrong
These costs follow legal duties, not how many regions are used.
Question 46 Play from 1:05:58
Tap an answer. The reason for that answer shows here.
Right, A is the answer.
Why A is right
With infrastructure as a service the provider supplies virtual hardware and the customer manages everything from the operating system upward. That gives the most control over the OS of any of the common service models.
Not this one. Try another answer.
Why B is wrong
The provider runs the application and its operating system, so the customer has almost no OS control.
Not this one. Try another answer.
Why C is wrong
Function platforms hide the operating system entirely and only run small pieces of code.
Not this one. Try another answer.
Why D is wrong
The provider manages the operating system and runtime so developers just deploy code.
Question 47 Play from 1:07:27
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
The attacker took advantage of the leak but did not cause it.
Not this one. Try another answer.
Why B is wrong
A repository is only a tool, and the customer decides what is stored in it.
Right, C is the answer.
Why C is right
Under shared responsibility, the provider secures the platform, but secrets and credentials belong to whoever writes and deploys the code. The platform was compliant and the leak came from how the firm handled its own keys, so the fault sits with the firm.
Not this one. Try another answer.
Why D is wrong
The platform was compliant, and the weakness was in the customer's handling of credentials.
Question 48 Play from 1:08:56
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
Retention is about how long data is kept, not who controls the keys.
Right, B is the answer.
Why B is right
Owning your data means deciding how it is protected, and that includes locking it with your own keys. Blocking those keys takes that control away from the customer.
Not this one. Try another answer.
Why C is wrong
Classification is about labeling data by sensitivity.
Not this one. Try another answer.
Why D is wrong
Sovereignty deals with which country's laws apply to the data's location.
Question 49 Play from 1:10:21
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
It exposes the change to a small slice of users first, which is slower and more involved.
Not this one. Try another answer.
Why B is wrong
It swaps instances in batches over time, so it is not the quick option.
Not this one. Try another answer.
Why C is wrong
It overwrites the running environment, adding risk and ignoring the second platform.
Right, D is the answer.
Why D is right
With two matching environments, one live and one idle, you push the new version to the idle side and then flip traffic over. That is simple and fast, and flipping back is just as easy if something breaks.
Question 50 Play from 1:11:45
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
It disappears once the container stops, so the results would be gone.
Right, B is the answer.
Why B is right
A persistent volume lives apart from the container, so files written there survive restarts and even deletion of the container. That makes it the right place for results you want to read afterward.
Not this one. Try another answer.
Why C is wrong
This is not a practical way to store container data in the cloud.
Not this one. Try another answer.
Why D is wrong
It is a text stream for the console or logs, not a place to keep files.
Question 51 Play from 1:13:01
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
Without load balancing, traffic cannot be steered away from a failed zone.
Not this one. Try another answer.
Why B is wrong
It only makes a single instance larger and leaves it in one zone.
Right, C is the answer.
Why C is right
When copies of the workload run in separate zones, the surviving zones keep serving users after one zone fails.
Not this one. Try another answer.
Why D is wrong
One server is a single point of failure.
Question 52 Play from 1:14:17
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
It follows a request across services and has nothing to do with storage duration.
Not this one. Try another answer.
Why B is wrong
It gathers logs from systems but does not decide how long they are kept.
Right, C is the answer.
Why C is right
The need here is about how long records are kept, so they can still be produced when auditors ask. Setting a keep-for-this-long rule on the logs covers exactly that.
Not this one. Try another answer.
Why D is wrong
It combines logs in one place without setting any keeping period.
Question 53 Play from 1:15:40
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
This is a human login detail, not an application credential.
Right, B is one of the answers.
Why B is right
Automation against a REST interface usually relies on an application identity pair: one value says who is calling and a private value proves it. Human login details and browser artifacts are not how scripts normally sign in.
Not this one. Try another answer.
Why C is wrong
Cookies belong to browser sessions, not scripted calls.
Not this one. Try another answer.
Why D is wrong
The name of a computer plays no part in API sign-in.
Right, E is one of the answers.
Why E is right
Automation against a REST interface usually relies on an application identity pair: one value says who is calling and a private value proves it. Human login details and browser artifacts are not how scripts normally sign in.
Not this one. Try another answer.
Why F is wrong
Passwords are for people, not for programmatic access.
Question 54 Play from 1:17:16
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
It controls where workloads are placed, not how they are reached.
Right, B is the answer.
Why B is right
Publishing a container's port on the host lets outside traffic arrive at the application running inside. Without that link, the service stays unreachable from outside.
Not this one. Try another answer.
Why C is wrong
It relays outbound client requests and does not publish a service.
Not this one. Try another answer.
Why D is wrong
It automates network setup in general but does not itself expose a container service.
Question 55 Play from 1:18:35
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
That describes a recurring service plan or payment arrangement.
Right, B is the answer.
Why B is right
A signed deal between two parties that sets out duties both must follow is a contract. Here the duty is encrypting shared data, and it binds only the two signers.
Not this one. Try another answer.
Why C is wrong
That refers to lawsuits and legal disputes, not a signed agreement.
Not this one. Try another answer.
Why D is wrong
Those rules come from governments or industry bodies, not from a deal between two firms.
Question 56 Play from 1:19:53
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
It measures how long a repair takes, not how much data disappears.
Right, B is the answer.
Why B is right
This target is measured as a span of time back from the failure and says how far the data may roll back. That span sets the most data you can afford to lose, which guides how often backups or replication must run.
Not this one. Try another answer.
Why C is wrong
It sets how long downtime may last, not the data loss.
Not this one. Try another answer.
Why D is wrong
It is the umbrella plan that contains the targets rather than the data-loss measure itself.
Question 57 Play from 1:21:18
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
It focuses on links between items, not general loose content.
Not this one. Try another answer.
Why B is wrong
It targets embeddings used for similarity search.
Right, C is the answer.
Why C is right
Databases without a fixed table design accept varied formats and let the structure evolve, which suits loose content that keeps changing.
Not this one. Try another answer.
Why D is wrong
It demands a fixed schema in tables, which fights against constant change.
Question 58 Play from 1:22:34
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
It is a legacy platform and not a modern cloud-native design.
Right, B is the answer.
Why B is right
Splitting an application into many small services, each deployed and updated on its own and talking through APIs, gives loose coupling and easy upkeep.
Not this one. Try another answer.
Why C is wrong
Everything sits in one tightly bound codebase, the opposite of decoupled.
Not this one. Try another answer.
Why D is wrong
It describes mixing on-premises and cloud hosting, not an application design.
Question 59 Play from 1:23:52
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
It is common on both solid-state and spinning drives.
Not this one. Try another answer.
Why B is wrong
It carries storage traffic over a network and can sit in front of either drive type.
Right, C is the answer.
Why C is right
This interface was built for flash storage and talks straight over PCIe, taking advantage of very low delay. In practice it is found on solid-state drives, while the rest are used with both drive types.
Not this one. Try another answer.
Why D is wrong
Enterprise systems use it with both kinds of drive.
Question 60 Play from 1:25:12
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
It deals with written and spoken language, not visuals.
Right, B is the answer.
Why B is right
This field covers anything a machine does to make sense of visual input, from spotting objects to following people on screen.
Not this one. Try another answer.
Why C is wrong
It is just one narrow use of the wider field.
Not this one. Try another answer.
Why D is wrong
It is a technique for rebuilding images from data, a smaller piece of the field.
Question 61 Play from 1:26:26
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
Latency is an outcome of how the storage performs, not an input to the IOPS sizing math.
Not this one. Try another answer.
Why B is wrong
Memory use has no bearing on how many storage operations a workload needs.
Right, C is the answer.
Why C is right
An IOPS figure only turns into real data movement once you know how large each operation is. Small blocks mean a great many operations to move a given amount of data, while large blocks need far fewer. That is why the sizing math cannot be done without knowing the size of each block.
Not this one. Try another answer.
Why D is wrong
Network capacity relates to bandwidth, not to the count of storage operations per second.
Question 62 Play from 1:27:56
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
Availability matters, but it has to be delivered within the legal limits, so it ranks below the mandated rules.
Right, B is one of the answers.
Why B is right
Public sector bodies answer to strict legal frameworks, so any cloud they pick has to satisfy mandated rules and standards such as FedRAMP. Those obligations come first and then shape every other choice in the move, which is why regulatory and compliance requirements top the list.
Not this one. Try another answer.
Why C is wrong
A schedule is a project planning detail that gets set once the mandatory requirements are known.
Right, D is one of the answers.
Why D is right
Public sector bodies answer to strict legal frameworks, so any cloud they pick has to satisfy mandated rules and standards such as FedRAMP. Those obligations come first and then shape every other choice in the move, which is why regulatory and compliance requirements top the list.
Not this one. Try another answer.
Why E is wrong
Firewall features are a design detail that follows from the rules the agency must obey.
Not this one. Try another answer.
Why F is wrong
Picking a service model is a later design decision, not the top concern for a regulated agency.
Question 63 Play from 1:29:46
Tap an answer. The reason for that answer shows here.
Right, A is the answer.
Why A is right
A web socket keeps one connection open so the client and server can send data to each other at any moment. Messaging tools rely on exactly that behavior, because new messages must show up instantly without the client asking over and over.
Not this one. Try another answer.
Why B is wrong
Moving files works fine over ordinary HTTP requests and does not need a constantly open two-way channel.
Not this one. Try another answer.
Why C is wrong
Dashboards usually refresh on a timer through standard API calls rather than needing a permanent socket.
Not this one. Try another answer.
Why D is wrong
Sharing pictures is typically handled with regular HTTP uploads and downloads.
Question 64 Play from 1:31:16
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
Looking up CVEs makes sense only after scan findings exist to look up.
Right, B is the answer.
Why B is right
Every vulnerability program begins by deciding exactly which assets and environments are in play. Until that boundary is set, there is nothing to scan, rank, or fix, so everything else depends on this opening move.
Not this one. Try another answer.
Why C is wrong
Fixing problems comes at the end, once they have been found and ranked.
Not this one. Try another answer.
Why D is wrong
Assessing systems needs a defined target list, so it follows the scoping step.
Question 65 Play from 1:32:37
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
Remediation is about applying fixes, which happens after the findings are reviewed.
Not this one. Try another answer.
Why B is wrong
Scope is decided up front to choose what gets scanned, not to review results.
Right, C is the answer.
Why C is right
This is the stage where raw scanner output gets examined, false alarms are separated from genuine issues, and the real ones are written down. Evaluating and validating results like this is the heart of the assessment phase, which sits between the scan itself and the fixes.
Not this one. Try another answer.
Why D is wrong
Identification is the discovery of possible issues, not the checking of findings for accuracy.
Question 66 Play from 1:34:10
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
Load-based scaling reacts only after traffic has already climbed, so users still feel the slowdown first.
Right, B is the answer.
Why B is right
Because the surges follow a predictable calendar, capacity can be added ahead of time on a timetable. Resources are already in place when viewers arrive, and nobody has to step in during the rush.
Not this one. Try another answer.
Why C is wrong
Manual scaling depends on a person acting at the right moment and uses none of the known timing.
Not this one. Try another answer.
Why D is wrong
Event-driven scaling waits for a trigger to fire instead of using the forecast release dates.
Question 67 Play from 1:35:35
Tap an answer. The reason for that answer shows here.
Right, A is the answer.
Why A is right
Mixing local infrastructure with services in a provider's cloud is the textbook definition of a hybrid design. Some workloads live in the company's own facility and others run remotely, and the two are used together.
Not this one. Try another answer.
Why B is wrong
A community cloud is shared by several groups with common needs, which is not the case here.
Not this one. Try another answer.
Why C is wrong
A public cloud means everything runs on the provider's shared platform, with nothing kept in-house.
Not this one. Try another answer.
Why D is wrong
A private cloud is dedicated to one organization and does not split workloads with an outside provider.
Question 68 Play from 1:37:02
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
Zero Trust is a design philosophy for access control, not a list of flaws.
Not this one. Try another answer.
Why B is wrong
CWE groups types of coding weaknesses in general and does not track individual known vulnerabilities.
Not this one. Try another answer.
Why C is wrong
Data security standards are compliance rules such as PCI DSS, not a vulnerability list.
Right, D is the answer.
Why D is right
CVE entries each describe one specific flaw that has been publicly disclosed and given its own tracking number. Security teams use that shared catalog to look up and talk about known problems in the same way.
Question 69 Play from 1:38:43
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
A SAN is a costly enterprise storage network and a poor match for a low-cost cloud move.
Right, B is the answer.
Why B is right
Data is kept as self-describing objects in a flat layout, which lets capacity grow almost without limit at a low price per gigabyte. That combination is why cloud teams default to it for large pools of unstructured content.
Not this one. Try another answer.
Why C is wrong
File storage relies on a folder hierarchy that scales less well and costs more for huge unstructured sets.
Not this one. Try another answer.
Why D is wrong
Block storage is meant for individual volumes, and its price per unit is higher for bulk content.
Question 70 Play from 1:40:15
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
Identification is the discovery of a problem, which would have happened before any blocking rule was written.
Not this one. Try another answer.
Why B is wrong
Assessment ranks and analyzes findings rather than putting a technical control in place.
Right, C is the answer.
Why C is right
Cutting off a hostile address with a firewall rule is a hands-on action that reduces or removes the risk. Taking a protective step like this is what remediation means, as opposed to finding, measuring, or planning what to scan.
Not this one. Try another answer.
Why D is wrong
Scope only defines which assets get scanned and does not involve filtering any traffic.
Question 71 Play from 1:41:48
Tap an answer. The reason for that answer shows here.
Right, A is the answer.
Why A is right
Checking integrity confirms that the saved data is complete and uncorrupted, and that a restore will bring it all back. Since the concern is a faster job possibly losing something, proving the copies are intact is the right test.
Not this one. Try another answer.
Why B is wrong
The schedule only sets when jobs run and says nothing about whether a restore is complete.
Not this one. Try another answer.
Why C is wrong
Encryption guards confidentiality but does not prove the data comes back whole.
Not this one. Try another answer.
Why D is wrong
Retention controls how long copies are kept, not whether they can be restored.
Question 72 Play from 1:43:15
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
A commit records changes in the local repository only and never contacts the server.
Not this one. Try another answer.
Why B is wrong
A pull downloads remote changes and merges them into the local branch, the opposite direction.
Right, C is the answer.
Why C is right
Pushing sends the commits you made on your machine up to the server copy, so teammates can see them. Every other choice either stays local or brings data down instead of sending it up.
Not this one. Try another answer.
Why D is wrong
A fetch only downloads remote updates and sends nothing back.
Question 73 Play from 1:44:32
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
Writing straight to main skips any review, so a bad setting can go live right away.
Not this one. Try another answer.
Why B is wrong
Editing in the browser sidesteps the normal review and testing workflow and raises the chance of errors.
Not this one. Try another answer.
Why C is wrong
A fetch merely downloads updates and adds no check on the quality of a change.
Right, D is the answer.
Why D is right
A pull request puts a change in front of other people before it is merged. A second pair of eyes tends to catch mistakes in a template early, long before the faulty setting could reach a live environment.
Question 74 Play from 1:46:05
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
A watermark marks ownership and does nothing about weaknesses in the image contents.
Not this one. Try another answer.
Why B is wrong
A firewall guards a running container from outside traffic, but the flaws inside the image remain.
Not this one. Try another answer.
Why C is wrong
A signature proves who published the image and that it was not altered, yet it does not remove any vulnerabilities.
Right, D is the answer.
Why D is right
Images built to Center for Internet Security benchmarks start out with unneeded components removed, patches applied, and safer defaults set. That shrinks the number of flaws inside the image itself, which is exactly the analyst's worry.
Question 75 Play from 1:47:45
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
Cold storage is cheaper than hot, but it still costs more per gigabyte than the archive tier.
Not this one. Try another answer.
Why B is wrong
Warm sits in the middle on price and speed, so it wastes money on data that is rarely touched.
Right, C is the answer.
Why C is right
Data that must be kept for years but is rarely read belongs in the tier with the lowest storage price. The trade-off is slower and costlier retrieval, which is acceptable for records that sit untouched for long periods.
Not this one. Try another answer.
Why D is wrong
Hot storage gives the fastest access at the highest price, which is unnecessary here.
Question 76 Play from 1:49:07
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
Block storage is a durable disk-style volume that survives restarts when attached.
Not this one. Try another answer.
Why B is wrong
Object storage is a durable service whose contents do not depend on a container's lifecycle.
Not this one. Try another answer.
Why C is wrong
A persistent volume exists precisely to keep data after a container restarts.
Right, D is the answer.
Why D is right
Ephemeral storage lives and dies with the container, so a restart wipes whatever was written there. To keep data across restarts it must be placed on storage that exists independently of the container.
Question 77 Play from 1:50:28
Tap an answer. The reason for that answer shows here.
Right, A is the answer.
Why A is right
Microservices split an application into small, independent pieces that each handle one job and talk through defined APIs. Because the pieces do not depend heavily on each other, teams can build, release, and scale them separately.
Not this one. Try another answer.
Why B is wrong
Microservices can run on a private cloud, but that describes where they run, not the design pattern.
Not this one. Try another answer.
Why C is wrong
Tight coupling is the reverse of this approach and fits a single monolithic application.
Not this one. Try another answer.
Why D is wrong
A public cloud is a hosting choice, while microservices describe how the software is structured.
Question 78 Play from 1:51:59
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
Tickets record that a change happened but do nothing to make deployments identical.
Right, B is the answer.
Why B is right
Writing the setup down as code means every deployment runs from the same definition, which can be versioned and reviewed. Nobody has to improvise, so the engineers get the same result each time and the drift between deployments goes away.
Not this one. Try another answer.
Why C is wrong
Logs show what occurred after the fact and cannot prevent differences while deploying.
Not this one. Try another answer.
Why D is wrong
Written instructions still depend on each person's interpretation, which is the cause of the problem.
Question 79 Play from 1:53:30
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
Hybrid blends several models and does not by itself deliver the cost savings of broad sharing.
Right, B is the answer.
Why B is right
In a public cloud, many customers run on the same pooled hardware, so the provider spreads the expense across all of them. That sharing is what drives the lowest prices of any model.
Not this one. Try another answer.
Why C is wrong
A private cloud is dedicated to one organization, so no cost is split with others.
Not this one. Try another answer.
Why D is wrong
A community cloud is shared only within a small group, which means fewer tenants and a higher cost per member.
Question 80 Play from 1:54:52
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
A port scan looks for open services and does nothing about leaked secrets.
Right, B is one of the answers.
Why B is right
Once secrets are visible in a repository, anyone could use them. The urgent move is to shut the account off so it cannot be abused, and then replace the secrets with new ones so the leaked copies are worthless.
Not this one. Try another answer.
Why C is wrong
Cipher settings concern encryption of traffic, not credentials that are already exposed.
Not this one. Try another answer.
Why D is wrong
Adding multifactor authentication does not undo the exposure of the secrets that already leaked.
Right, E is one of the answers.
Why E is right
Once secrets are visible in a repository, anyone could use them. The urgent move is to shut the account off so it cannot be abused, and then replace the secrets with new ones so the leaked copies are worthless.
Not this one. Try another answer.
Why F is wrong
A VPN does not resolve the underlying problem of exposed service account credentials.
Question 81 Play from 1:56:35
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
Git only tracks changes to files and does not deploy anything.
Not this one. Try another answer.
Why B is wrong
Jenkins runs build and release pipelines, but it does not provision infrastructure from templates.
Right, C is the answer.
Why C is right
The clues are template-driven deployment and a hybrid target, which points to infrastructure as code. Terraform reads reusable definitions and talks to many providers, so one workflow can build out both a private data center and a public cloud. The other tools handle different parts of the software lifecycle.
Not this one. Try another answer.
Why D is wrong
Docker packages an app into a container, yet it has no built-in way to provision environments across providers.
Question 82 Play from 1:58:08
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
Uptime is a quality measure and says nothing about how easily you can leave.
Not this one. Try another answer.
Why B is wrong
Price is a business concern, not a guard against being tied to one vendor.
Right, C is one of the answers.
Why C is right
Lock-in happens when your tools and your data are tied to a single provider. Building on open standards keeps your choices open, and being able to move your data elsewhere with little effort means you can leave if you need to. The remaining factors matter when picking a vendor, but they do not protect you from getting stuck.
Not this one. Try another answer.
Why D is wrong
Compliance is a requirement to meet, but it applies no matter which vendor you pick.
Not this one. Try another answer.
Why E is wrong
Storage features describe what a vendor offers, not how portable your workloads are.
Right, F is one of the answers.
Why F is right
Lock-in happens when your tools and your data are tied to a single provider. Building on open standards keeps your choices open, and being able to move your data elsewhere with little effort means you can leave if you need to. The remaining factors matter when picking a vendor, but they do not protect you from getting stuck.
Question 83 Play from 1:59:57
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
Integrity is about data not being altered improperly, not about how long it is kept.
Not this one. Try another answer.
Why B is wrong
Encryption protects confidentiality and does not set a storage period.
Not this one. Try another answer.
Why C is wrong
Recoverability is the ability to restore systems or data, not a required keeping time.
Right, D is the answer.
Why D is right
When a legal team says to keep data for a set length of time, that is a rule about how long records must be preserved. That is what retention means, and it applies regardless of how the incident itself was handled.
Question 84 Play from 2:01:21
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
Text recognition pulls characters out of images and does not hold a conversation.
Not this one. Try another answer.
Why B is wrong
A help page is static content that visitors read, not an interactive exchange.
Right, C is the answer.
Why C is right
A conversational assistant that understands plain-language questions and replies on the spot gives visitors an interactive way to fix problems. Generative AI is the technology that powers that kind of experience. The other choices only show fixed content or convert images to text.
Not this one. Try another answer.
Why D is wrong
A search engine returns links or documents rather than conversing to solve a problem.
Question 85 Play from 2:02:51
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
Testing checks that code works, it does not record who changed what.
Right, B is the answer.
Why B is right
Tracking every edit to source code, with the ability to browse history and roll back, is called versioning, and tools such as Git are built for exactly that. The other terms either describe unrelated activities or are not standard practices.
Not this one. Try another answer.
Why C is wrong
Drift describes configurations slowly diverging from their intended state, not a way to log changes.
Not this one. Try another answer.
Why D is wrong
This is not a recognized practice for keeping a history of changes.
Question 86 Play from 2:04:16
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
A VM must be kept running and patched even when the script is idle.
Right, B is the answer.
Why B is right
A job that runs one short script on a timer does not need a machine running all day. A serverless function starts only when triggered, bills for the brief time it runs, and needs no server upkeep. The other options leave you paying for and maintaining resources that sit idle most of the time.
Not this one. Try another answer.
Why C is wrong
Dedicating physical hardware to one small script is costly and wasteful.
Not this one. Try another answer.
Why D is wrong
A container platform adds orchestration overhead that a single scheduled script does not need.
Question 87 Play from 2:05:44
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
IaaS leaves the customer managing operating systems and applications on rented infrastructure.
Not this one. Try another answer.
Why B is wrong
FaaS runs short event-driven code, not a full business application.
Right, C is the answer.
Why C is right
If the goal is minimal administration, the provider should run as much of the stack as possible. With software as a service, the vendor handles servers, the platform, updates and security, and the company simply uses the CRM. The other models leave the company managing at least the application itself, or are meant for small event-driven functions.
Not this one. Try another answer.
Why D is wrong
PaaS still requires the company to build and manage its own application.
Question 88 Play from 2:07:14
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
Mining is not storage heavy, so disk activity does not stand out.
Right, B is the answer.
Why B is right
Mining digital coins is computation heavy, so hijacked instances show processor usage pinned high for long stretches with no matching business workload. Watching that percentage is the simplest way to catch it. Other metrics may move too, but they are not the telltale sign.
Not this one. Try another answer.
Why C is wrong
Memory use stays fairly modest during mining and is not a reliable signal.
Not this one. Try another answer.
Why D is wrong
Mining traffic is light, so packet counts do not reveal it clearly.
Question 89 Play from 2:08:44
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
Access control lists limit who can reach resources, but they do not tighten the image contents.
Not this one. Try another answer.
Why B is wrong
Least privilege governs permissions granted to accounts, not the baseline build of the image.
Not this one. Try another answer.
Why C is wrong
Scanning only finds problems and does not remove or reconfigure anything.
Right, D is the answer.
Why D is right
Starting from a public image means you inherit unknown settings, extra services and possible weaknesses. Hardening addresses this by turning off what is not needed, closing ports, patching and tightening configuration before the image becomes the company standard. The other choices are useful controls but do not by themselves make the image itself secure.
Question 90 Play from 2:10:15
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
NFS shares files over a network and is not designed for lightweight device messaging.
Not this one. Try another answer.
Why B is wrong
HTTPS is a general web protocol with heavier overhead per message.
Not this one. Try another answer.
Why C is wrong
SMB is also a file sharing protocol, not an IoT messaging one.
Right, D is the answer.
Why D is right
Small devices often run on weak links and limited power, so the protocol needs a tiny header and a publish and subscribe design. MQTT is built for that, which keeps the amount of data sent low. File sharing protocols and general web traffic carry far more overhead.
Question 91 Play from 2:11:48
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
Shadow deployments mirror production traffic to the new version, which increases load.
Right, B is the answer.
Why B is right
A rolling release swaps old instances for new ones in small batches, so extra capacity is never needed all at once and traffic stays steady. Blue-green runs a full second environment and shadow testing duplicates live traffic, so both add load. A/B testing is about comparing user behavior, not about a safe rollout.
Not this one. Try another answer.
Why C is wrong
A/B testing compares versions on users to measure results, rather than rolling out safely.
Not this one. Try another answer.
Why D is wrong
Blue-green needs a complete duplicate environment, consuming extra resources.
Question 92 Play from 2:13:21
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
This is a certifiable management system standard, not an auditor's report on service controls.
Not this one. Try another answer.
Why B is wrong
This is a framework for IT service management, not a data protection audit.
Not this one. Try another answer.
Why C is wrong
These are hardening benchmarks for configuring systems, not a provider audit procedure.
Right, D is the answer.
Why D is right
A SOC 2 report is produced by an independent auditor and describes how a service provider manages customer data against trust criteria such as security and privacy. Customers often request it when evaluating a provider.
Question 93 Play from 2:14:52
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
Phishing tricks people into giving up credentials and does not slow a site.
Right, B is the answer.
Why B is right
Slow responses on a public page when nothing changed inside suggests the cause is outside: a flood of requests from many sources swamping the service. That is a distributed denial of service attack. The others do not produce sudden slowness for users worldwide.
Not this one. Try another answer.
Why C is wrong
A staff mistake would appear as an internal change or misconfiguration, which monitoring would have caught.
Not this one. Try another answer.
Why D is wrong
Cryptojacking would show up as high CPU on the hosts, which internal monitoring would flag.
Question 94 Play from 2:16:27
Tap an answer. The reason for that answer shows here.
Right, A is the answer.
Why A is right
Grafana is a dashboard tool that charts metrics and logs, so people open it to see how systems are performing. It is not used to build containers, ship releases or store code.
Not this one. Try another answer.
Why B is wrong
Containers are built with tools like Docker, not Grafana.
Not this one. Try another answer.
Why C is wrong
Releases go through deployment or CI/CD tools.
Not this one. Try another answer.
Why D is wrong
Code is pushed with Git clients, not Grafana.
Question 95 Play from 2:17:50
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
An availability zone provides fault isolation within a region, not proximity for content.
Right, B is the answer.
Why B is right
To cut latency, content is placed physically near users, which is the job of edge locations. Regions and availability zones are broad hosting locations for workloads, not delivery points close to users, and bursting is about overflow capacity.
Not this one. Try another answer.
Why C is wrong
A region is a large geographic area for hosting, too coarse to serve content near each user.
Not this one. Try another answer.
Why D is wrong
Cloud bursting shifts overflow demand to a public cloud and is not about location.
Question 96 Play from 2:19:18
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
XaaS is a general label for anything delivered as a service, not a specific model for hosting code.
Not this one. Try another answer.
Why B is wrong
IaaS requires provisioning and managing virtual machines and networks.
Not this one. Try another answer.
Why C is wrong
SaaS delivers a finished application and does not host your own code.
Right, D is the answer.
Why D is right
When a team only wants to push its code and let the provider handle servers, storage and networking, a platform service is the match. Developers focus entirely on the application. Software services are finished products, infrastructure services require setting up machines, and XaaS is just a broad umbrella term.
Question 97 Play from 2:20:54
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
MFA requires a person to respond to a prompt, which a service cannot do.
Not this one. Try another answer.
Why B is wrong
Cookies keep browser sessions alive and are not a credential for an application.
Right, C is the answer.
Why C is right
Applications need a non-interactive way to prove who they are. An API key can be issued per application, limited in scope, and rotated, so it suits this job. Methods that rely on a human being present are a poor fit.
Not this one. Try another answer.
Why D is wrong
Human-style credentials are harder to scope and rotate and are not the best practice for services.
Question 98 Play from 2:22:22
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
Rebasing rewrites history on that branch, which protection typically blocks too.
Not this one. Try another answer.
Why B is wrong
A direct merge is still a direct change to the protected branch, so it is also blocked.
Not this one. Try another answer.
Why C is wrong
A new remote only adds a repository location and does not get around the restriction.
Right, D is the answer.
Why D is right
Protected branches are locked so nobody can change them directly. The expected route is to put the changes on a separate branch and open a pull request, which lets teammates review and then merge through the rules. The alternatives still try to alter or sidestep the protection.
Question 99 Play from 2:23:56
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
Upgrading an OS is an infrastructure change, not a code change.
Not this one. Try another answer.
Why B is wrong
A registry is a storage service for images and involves no change to the application's code.
Not this one. Try another answer.
Why C is wrong
Changing file formats concerns data, not application code.
Right, D is the answer.
Why D is right
Refactoring means changing an application's code so it works well on a new target, here a hosting platform. The other choices describe infrastructure, operating system, or file-level tasks that leave the application code alone.
Question 100 Play from 2:25:23
Tap an answer. The reason for that answer shows here.
Not this one. Try another answer.
Why A is wrong
A load balancer spreads requests across servers but is not a centrally programmable network fabric.
Not this one. Try another answer.
Why B is wrong
A VPN builds encrypted tunnels and does not adapt to traffic.
Not this one. Try another answer.
Why C is wrong
A CDN caches content near users and does not provide central network control.
Right, D is the answer.
Why D is right
Centralized control, programmable behavior, and automatic adjustment to traffic are the hallmarks of software-defined networking, where a controller manages the network through software. Load balancers, VPNs and content delivery networks each handle a narrower task.
Transcript
Play from 00:00 Question 1. Several independent bookstores in one region want to share the cost of moving their servers to the cloud together. Which deployment model suits this group best? Here are the options. Option A. Private. Option B. Community. Option C. Public. Option D. Hybrid. Pause the video now if you want time to think about the answer.
The correct answer is B. Community. When a handful of organizations in the same line of business have similar needs, they can split the bill for one shared environment that only they use. That gives them savings from pooling plus more control than a general-purpose public offering. Now, why are the other options incorrect? Option A is incorrect. Each bookstore would pay for its own dedicated setup, which drives the price up. Option C is incorrect. It is open to everyone and gives this group no special shared arrangement among themselves. Option D is incorrect. It mixes on-premises and cloud resources, which adds complexity the group does not need. Therefore, B, Community, is the correct answer.
Play from 01:29 Question 2. An operations team sees a web tier that keeps climbing above roughly 75 to 80 percent utilization. Which scaling method should react to that condition? Here are the options. Option A. Scheduled. Option B. Load. Option C. Manual. Option D. Trending. Pause the video now if you want time to think about the answer.
The correct answer is B. Load. Scaling based on load watches live utilization and adds capacity the moment a threshold is crossed. That is exactly what you want when usage rises past a set percentage, with no person or calendar involved. Now, why are the other options incorrect? Option A is incorrect. It follows a calendar, so it cannot react to an unexpected jump in usage. Option C is incorrect. It depends on someone noticing and acting, which is slow. Option D is incorrect. It looks at past patterns to predict demand rather than reacting to current utilization. Therefore, B, Load, is the correct answer.
Play from 02:51 Question 3. A startup wants to get its code running with the least setup effort and the fastest start. Which compute option fits best? Here are the options. Option A. Snapshot. Option B. Container. Option C. Cluster. Option D. Serverless. Pause the video now if you want time to think about the answer.
The correct answer is D. Serverless. With this model you hand over your code and the provider takes care of servers, scaling, and capacity for you. There is almost nothing to configure, so it is the quickest path from idea to running service. Now, why are the other options incorrect? Option A is incorrect. It is a point-in-time copy used for recovery, not a way to run workloads. Option B is incorrect. You still have to build images, store them in a registry, and usually run an orchestrator. Option C is incorrect. You must set up nodes and orchestration before anything runs. Therefore, D, Serverless, is the correct answer.
Show full transcriptShow less
Play from 04:09 Question 4. A records department wants software that can pull the words off a hard-copy notice. Which technology does that job? Here are the options. Option A. Natural language processing. Option B. Text recognition. Option C. Sentiment analysis. Option D. Document scanning. Pause the video now if you want time to think about the answer.
The correct answer is B. Text recognition. This technology, often called OCR, looks at the image of paper and converts the characters into text a computer can use. The other choices either only capture a picture or work on text that is already digital. Now, why are the other options incorrect? Option A is incorrect. It interprets text that is already in digital form and does not read paper. Option C is incorrect. It judges the mood or opinion in text rather than reading it from paper. Option D is incorrect. It produces an image of the page but does not pull out the words. Therefore, B, Text recognition, is the correct answer.
Play from 05:32 Question 5. A gaming company is publishing a service that needs the fastest possible connection handling with minimal delay. Which service should front it? Here are the options. Option A. Application gateway. Option B. Virtual private cloud. Option C. Network load balancer. Option D. Application load balancer. Pause the video now if you want time to think about the answer.
The correct answer is C. Network load balancer. A load balancer working at the transport layer passes traffic along without inspecting the content of each request. Because of that light touch it adds very little delay and can handle huge connection volumes. Now, why are the other options incorrect? Option A is incorrect. It examines application-level content, which adds processing time. Option B is incorrect. It is an isolated network space, not a service for distributing and publishing traffic. Option D is incorrect. It makes decisions on application data, so it adds more delay than a transport-layer option. Therefore, C, Network load balancer, is the correct answer.
Play from 06:57 Question 6. In a software team's repository, the production build moves from version 3.2 to version 3.3. What kind of release is this? Here are the options. Option A. Beta. Option B. Major. Option C. Alpha. Option D. Minor. Pause the video now if you want time to think about the answer.
The correct answer is D. Minor. Under the common numbering scheme, bumping the second number signals new capabilities that still work with existing users. A jump in the first number would mark a major change instead, and the pre-release labels apply to unfinished software. Now, why are the other options incorrect? Option A is incorrect. It is a test-stage pre-release, not a production update. Option B is incorrect. That would show up as a change in the first number, such as 3.2 to 4.0. Option C is incorrect. It is an early, unfinished build and not a production release. Therefore, D, Minor, is the correct answer.
Play from 08:17 Question 7. A retailer is hosting a classic SQL database and needs the storage type that gives it the best speed. Which type should it pick? Here are the options. Option A. Ephemeral. Option B. Block. Option C. Object. Option D. File. Pause the video now if you want time to think about the answer.
The correct answer is B. Block. Databases do a lot of small random reads and writes and need quick, direct access to disk-like volumes. Block volumes attach to a server and deliver that low delay and steady throughput. Now, why are the other options incorrect? Option A is incorrect. It disappears when the instance stops, so it cannot safely hold database data. Option C is incorrect. It suits large unstructured files and has higher latency, which hurts database work. Option D is incorrect. It is built for shared access to files, not for the random access patterns of a database. Therefore, B, Block, is the correct answer.
Play from 09:36 Question 8. An IT group wants to get more from its cloud spend and cut the work of patching many separate guest operating systems. Which compute choice helps most? Here are the options. Option A. Bare-metal servers. Option B. VM. Option C. Containers. Option D. Remote desktops. Pause the video now if you want time to think about the answer.
The correct answer is C. Containers. Containers share the host's operating system kernel, so many workloads can run without each one needing its own full OS. That saves resources and leaves far fewer systems to patch and maintain. Now, why are the other options incorrect? Option A is incorrect. Each machine still needs its own OS installed and maintained. Option B is incorrect. Every virtual machine carries its own guest OS that someone has to look after. Option D is incorrect. They add more operating system instances to manage rather than fewer. Therefore, C, Containers, is the correct answer.
Play from 10:56 Question 9. A payment processor's auditor asks which standard forbids sending or keeping card numbers as plain readable text. What is the answer? Here are the options. Option A. PCI-DSS. Option B. SOC2. Option C. GDPR. Option D. CSA. Pause the video now if you want time to think about the answer.
The correct answer is A. PCI-DSS. This standard was created by the card brands and sets rules for anyone handling cardholder data, including protecting it with encryption while stored and while moving across networks. Now, why are the other options incorrect? Option B is incorrect. It is an audit report on service provider controls, not a card data rulebook. Option C is incorrect. It is a broad European privacy law and is not specific to card numbers. Option D is incorrect. It publishes general cloud security guidance and does not set card data rules. Therefore, A, PCI-DSS, is the correct answer.
Play from 12:24 Question 10. Which term describes the rule that information must remain inside particular geographic or national boundaries? Here are the options. Option A. Data ownership. Option B. Data sovereignty. Option C. Data retention. Option D. Data classification. Pause the video now if you want time to think about the answer.
The correct answer is B. Data sovereignty. This idea says data falls under the laws of the place where it is stored or collected, so it often has to stay in that country or region. It is a legal matter tied to location. Now, why are the other options incorrect? Option A is incorrect. It concerns who holds rights and responsibility for the data, not where it sits. Option C is incorrect. It is about how long data is kept before being removed. Option D is incorrect. It sorts data by sensitivity level, not by geography. Therefore, B, Data sovereignty, is the correct answer.
Play from 13:43 Question 11. A retailer uses a managed database service from a cloud provider. Which two tasks stay with the retailer? (Choose two.) Here are the options. Option A. Row-level encryption. Option B. Operating system patches. Option C. Having spare hardware ready for growth. Option D. Table-level permissions. Option E. Small updates to the database engine. Option F. Cluster configuration. Pause the video now if you want time to think about the answer.
The correct answers are A. Row-level encryption, and D. Table-level permissions. With a managed database, the provider runs the underlying platform, while the customer still decides who can see which data and how sensitive records are protected. Access rules on tables and encryption of individual rows both fall on the customer side. Now, why are the other options incorrect? Option B is incorrect. The provider maintains the operating system under the managed service. Option C is incorrect. Supplying and scaling physical capacity is the provider's job. Option E is incorrect. The provider applies engine updates as part of the managed offering. Option F is incorrect. The provider sets up and runs the cluster behind the service. Therefore, A and D are the correct answers.
Play from 15:23 Question 12. A software vendor ships one product to many customers, each getting a different set of features, and must keep their code changes organized separately. Which version control feature fits? Here are the options. Option A. Merge. Option B. Push. Option C. Branch. Option D. Commit. Pause the video now if you want time to think about the answer.
The correct answer is C. Branch. Creating a separate line of development lets each customer's variation evolve on its own while still starting from the shared product code. Changes can later be brought back together when appropriate. Now, why are the other options incorrect? Option A is incorrect. It joins work from different lines together but does not create the separate tracks. Option B is incorrect. It only sends commits to a remote repository. Option D is incorrect. It saves a set of changes but does not separate work per customer. Therefore, C, Branch, is the correct answer.
Play from 16:42 Question 13. Which storage type is the usual foundation for a data lake built on a public cloud provider? Here are the options. Option A. File. Option B. Ephemeral. Option C. Block. Option D. Object. Pause the video now if you want time to think about the answer.
The correct answer is D. Object. Data lakes hold enormous amounts of mixed, mostly unstructured data. Object storage scales almost without limit, costs less per unit, is durable, and plugs easily into analytics tools. Now, why are the other options incorrect? Option A is incorrect. Its hierarchical design does not scale or save money as well at data lake size. Option B is incorrect. It is temporary and disappears, so it cannot hold data that must be kept. Option C is incorrect. It fits databases and servers needing fixed volumes, not massive unstructured collections. Therefore, D, Object, is the correct answer.
Play from 18:00 Question 14. A retailer finds that its payment app still relies on an outdated Apache Log4j library. Which scoring system should the team use to rate how severe that exposure is? Here are the options. Option A. CVE. Option B. CWSS. Option C. CVSS. Option D. CWE. Pause the video now if you want time to think about the answer.
The correct answer is C. CVSS. This system turns the traits of a known vulnerability into a number, which helps teams decide what to fix first. The other acronyms name or categorize problems rather than rating how bad a specific flaw is. Now, why are the other options incorrect? Option A is incorrect. It gives vulnerabilities an identifier and a catalog entry but no severity score. Option B is incorrect. It scores types of software weaknesses, not a vulnerability in a running system. Option D is incorrect. It is a list of weakness categories and has no severity calculation. Therefore, C, CVSS, is the correct answer.
Play from 19:28 Question 15. A regional hospital group asks a managed service provider to help move its data center to the cloud. Which two considerations should the provider weigh most heavily? (Choose two.) Here are the options. Option A. Compute. Option B. Compliance. Option C. Cost. Option D. Security. Option E. Storage. Option F. Availability. Pause the video now if you want time to think about the answer.
The correct answers are B. Compliance, and D. Security. Health organizations handle private patient records and must follow strict regulations. That makes protecting the data and meeting legal obligations the top concerns, ahead of the technical sizing choices that follow from them. Now, why are the other options incorrect? Option A is incorrect. Processing capacity is a sizing detail that follows from the security and legal needs. Option C is incorrect. Budget matters, but it ranks below protecting patient data and meeting regulations. Option E is incorrect. Capacity is planned after the security and compliance requirements are set. Option F is incorrect. Uptime is valuable but is not the leading concern compared with protection and regulation. Therefore, B and D are the correct answers.
Play from 21:10 Question 16. A manufacturer wants to stress test its homegrown application for four weeks on a fleet of well over a hundred servers. Which service model is the best fit? Here are the options. Option A. SaaS. Option B. IaaS. Option C. DBaaS. Option D. PaaS. Pause the video now if you want time to think about the answer.
The correct answer is B. IaaS. Renting raw virtual servers by the hour lets the team spin up a large fleet for a short test and release it afterward. They keep full control of what runs on those machines and pay no long-term commitment. Now, why are the other options incorrect? Option A is incorrect. It delivers finished software and gives no way to deploy their own code. Option C is incorrect. It only provides managed databases, not general servers. Option D is incorrect. It hides the servers, which limits the control needed to build a custom large cluster. Therefore, B, IaaS, is the correct answer.
Play from 22:30 Question 17. A sysadmin is installing software on an Ubuntu server. Which package manager would typically be used? Here are the options. Option A. DNF. Option B. APT. Option C. YUM. Option D. Pacman. Pause the video now if you want time to think about the answer.
The correct answer is B. APT. Ubuntu comes from the Debian family, and that family relies on this tool to install, update, and remove software. Now, why are the other options incorrect? Option A is incorrect. It belongs to Fedora and newer Red Hat style systems. Option C is incorrect. It is used on older Red Hat style distributions. Option D is incorrect. It is the package tool for Arch Linux. Therefore, B, APT, is the correct answer.
Play from 23:37 Question 18. A platform lead wants the prod, test, and dev setups to stay identical every time they are built. Which approach best achieves that? Here are the options. Option A. Placing separate Jenkins build agents in every environment. Option B. Installing the ELK logging stack on each environment. Option C. Defining infrastructure in Terraform code reused with per-environment variables. Option D. Deploying a Grafana dashboard in every environment. Pause the video now if you want time to think about the answer.
The correct answer is C. Defining infrastructure in Terraform code reused with per-environment variables. Writing the infrastructure once as code and feeding it different variable values means every environment is created from the same blueprint. Only intended settings such as size or count differ, which removes drift between them. Now, why are the other options incorrect? Option A is incorrect. Build agents run pipeline jobs and do not define or enforce how infrastructure is built. Option B is incorrect. It gathers and searches logs but does not provision infrastructure. Option D is incorrect. It visualizes metrics and has no role in making environments match. Therefore, C, Defining infrastructure in Terraform code reused with per-environment variables, is the correct answer.
Play from 25:21 Question 19. A hosting firm runs many sites on a pool of cloned virtual machines. Which service should sit in front to route incoming requests? Here are the options. Option A. Application-layer load balancer. Option B. Web application firewall. Option C. Forward proxy. Option D. Content delivery network. Pause the video now if you want time to think about the answer.
The correct answer is A. Application-layer load balancer. An application-layer balancer can read details such as the hostname and URL path, so it can send each request to the right backend while spreading the work across the pool of machines. Now, why are the other options incorrect? Option B is incorrect. It filters malicious traffic but does not distribute requests across servers. Option C is incorrect. It acts for clients going outward, not for servers receiving inbound traffic. Option D is incorrect. It caches content near users and does not route among backend virtual machines. Therefore, A, Application-layer load balancer, is the correct answer.
Play from 26:43 Question 20. A security architect must ensure every user is verified, is permitted, connects from a sanctioned location, and uses a company-issued device. Which approach does this best? Here are the options. Option A. Defense in depth. Option B. Zero Trust. Option C. Secure by design. Option D. Least-privilege access. Pause the video now if you want time to think about the answer.
The correct answer is B. Zero Trust. This model assumes nothing is safe by default and checks identity, permissions, location, and device status on every access attempt. That lines up with each of the listed conditions. Now, why are the other options incorrect? Option A is incorrect. It stacks multiple layers of protection but is a general strategy, not a per-access verification model. Option C is incorrect. It is about building security into a system from the start, not checking users and devices. Option D is incorrect. It limits permissions to what is needed but says nothing about location or device checks. Therefore, B, Zero Trust, is the correct answer.
Play from 28:12 Question 21. A systems administrator at a retail chain is planning the month-end backup job. The stored backup should stay as small as possible and the job should finish quickly. Which backup type is the best fit? Here are the options. Option A. Differential. Option B. Incremental. Option C. Tape. Option D. Full. Pause the video now if you want time to think about the answer.
The correct answer is B. Incremental. The incremental approach copies only what changed since the previous backup of any kind, so every run is tiny and quick. A differential job keeps growing because it recopies everything since the last full one. Full jobs copy the whole data set each time, and tape is a storage medium rather than a backup method. Now, why are the other options incorrect? Option A is incorrect. It recopies every change since the last full backup, so it grows larger and runs longer than an incremental job. Option C is incorrect. Tape is a storage medium, not a backup type, and it is usually slower. Option D is incorrect. It copies the entire data set on every run, which is the largest and slowest choice. Therefore, B, Incremental, is the correct answer.
Play from 29:44 Question 22. A media company hosts its web platform on virtual machines in a public cloud, and attackers keep flooding it with traffic from many sources. Which data source should the cloud engineer watch to spot these floods? Here are the options. Option A. Instance syslog. Option B. Host-based EDR agent logs. Option C. Network flow logs. Option D. Provider logs of account and API activity. Pause the video now if you want time to think about the answer.
The correct answer is C. Network flow logs. Flow logs record who is talking to whom, on which ports, and how much data moves. A flood of requests from many sources shows up there as a sharp, unusual jump in volume. That makes them the natural place to spot a distributed denial-of-service attack. Now, why are the other options incorrect? Option A is incorrect. It records events inside one server, so it cannot show the overall traffic pattern of an attack. Option B is incorrect. These track malware and suspicious behavior on hosts, not floods of network traffic. Option D is incorrect. They record account and API actions, not the packets hitting the workload. Therefore, C, Network flow logs, is the correct answer.
Play from 31:23 Question 23. A consultant is advising a logistics firm whose old in-house app cannot keep up with user load and costs a lot to maintain. Which migration approach is the best way to modernize it? Here are the options. Option A. Refactor. Option B. Replatform. Option C. Retain. Option D. Rehost. Pause the video now if you want time to think about the answer.
The correct answer is A. Refactor. The app is both too slow to scale and costly to keep alive, so it needs deeper changes than a simple move. Refactoring reworks the code and design for scale and easier upkeep, which solves both complaints. Now, why are the other options incorrect? Option B is incorrect. It swaps some platform pieces but keeps the same core design, so scaling and upkeep problems can remain. Option C is incorrect. Leaving the app as it is does nothing for demand or maintenance cost. Option D is incorrect. A lift-and-shift copies the app to new infrastructure unchanged, so the same limits and costs come along. Therefore, A, Refactor, is the correct answer.
Play from 32:53 Question 24. A records department keeps closed case files that are never edited and are pulled only on request. Which storage tier fits this data? Here are the options. Option A. Hot. Option B. Archive. Option C. Warm. Option D. Cold. Pause the video now if you want time to think about the answer.
The correct answer is B. Archive. The archive tier is built for long-term records that are written once and fetched only when someone asks, at the lowest cost per gigabyte. Hot data is used constantly and warm data now and then, while cold data is infrequent but is not defined as read-only retention. Now, why are the other options incorrect? Option A is incorrect. It serves data that is used all the time with fast access. Option C is incorrect. It holds data touched occasionally and is not defined by being read-only. Option D is incorrect. It is for infrequent access, but it is not defined as read-only retention storage. Therefore, B, Archive, is the correct answer.
Play from 34:17 Question 25. The pager goes off for a senior cloud engineer at a retailer: a critical alert says the storefront web server has stopped responding. What should happen first? Here are the options. Option A. Remediation. Option B. Escalation. Option C. Monitoring. Option D. Triage. Pause the video now if you want time to think about the answer.
The correct answer is D. Triage. The first move on a critical page is to size up the situation: how bad it is, how many users are hit, and what is likely causing it. That assessment is triage, and it guides every step after it, including whether to fix things directly or call for help. Now, why are the other options incorrect? Option A is incorrect. A fix cannot be chosen sensibly until the problem is understood. Option B is incorrect. Calling others in comes after a first assessment shows it is needed. Option C is incorrect. Tuning monitoring is a follow-up task, not the immediate reaction to an outage. Therefore, D, Triage, is the correct answer.
Play from 35:44 Question 26. A cloud engineer needs billing reports that show spending broken out by individual resource. What should be set up? Here are the options. Option A. Rightsizing. Option B. Tagging. Option C. Invoicing. Option D. Reserve instances. Pause the video now if you want time to think about the answer.
The correct answer is B. Tagging. Labels attached to individual resources, such as project or owner, flow into billing data so costs can be filtered and grouped per resource. Rightsizing and reserved purchasing are ways to save money later, and an invoice only shows totals. Now, why are the other options incorrect? Option A is incorrect. It adjusts resource sizes to save money but does not label costs. Option C is incorrect. An invoice summarizes charges overall without breaking them down by resource. Option D is incorrect. They give a discount for a usage commitment and add no per-resource cost detail. Therefore, B, Tagging, is the correct answer.
Play from 37:07 Question 27. A network architect must tie several VPCs and the head office data center into one hub. Which component does this? Here are the options. Option A. Transit gateway. Option B. Subnet. Option C. CDN. Option D. Peering. Pause the video now if you want time to think about the answer.
The correct answer is A. Transit gateway. A transit gateway acts as a central hub that many VPCs and an on-premises link can all plug into, and it passes traffic between them. Peering only joins two VPCs directly, a subnet lives inside one VPC, and a CDN is for caching content near users. Now, why are the other options incorrect? Option B is incorrect. It is just a slice of address space inside a single VPC. Option C is incorrect. It speeds up content delivery and does not link private networks. Option D is incorrect. It joins two VPCs directly and does not provide hub routing across many networks or to on-premises. Therefore, A, Transit gateway, is the correct answer.
Play from 38:35 Question 28. A defense contractor is setting up a cloud platform that stays inside its own facility and is limited to staff with a need-to-know. Which deployment model fits best? Here are the options. Option A. Public. Option B. Community. Option C. Hybrid. Option D. Private. Pause the video now if you want time to think about the answer.
The correct answer is D. Private. A private cloud serves one organization only, and it can sit inside its own facilities with tightly controlled need-to-know permissions. Sharing with other organizations, opening to the internet, or mixing in public resources all weaken that control. Now, why are the other options incorrect? Option A is incorrect. It is open to anyone, so need-to-know limits cannot be guaranteed. Option B is incorrect. It is shared by several organizations with common interests, which widens access. Option C is incorrect. It blends in public cloud resources, which brings in outside exposure. Therefore, D, Private, is the correct answer.
Play from 40:01 Question 29. A regional insurer's auditors tell the backup administrator that backups must now support long-term reporting. Which change matters most? Here are the options. Option A. How long the backups are kept. Option B. Where the backups are stored. Option C. Which kind of backup is taken. Option D. How often the backups run. Pause the video now if you want time to think about the answer.
The correct answer is A. How long the backups are kept. Reports that look far back in time need the data to still exist, so the length of time backups are kept is the key setting. Everything else tunes how backups are made or stored, not how long they survive. Now, why are the other options incorrect? Option B is incorrect. Storing backups elsewhere changes where data sits, not how long it remains available. Option C is incorrect. Full, incremental, or differential affects speed and size, not how long data is preserved. Option D is incorrect. Frequency decides how many restore points exist, not how long they are kept. Therefore, A, How long the backups are kept, is the correct answer.
Play from 41:33 Question 30. A hosting provider wants to hand each customer an isolated compute unit with its own virtual network, operating system, CPU, and memory, carved out of shared physical hardware. Which option describes this? Here are the options. Option A. Container. Option B. Clustering. Option C. Stand-alone. Option D. Virtual machine. Pause the video now if you want time to think about the answer.
The correct answer is D. Virtual machine. A virtual machine emulates a full computer on top of physical hardware, with its own virtual network, operating system, processor share, and memory. That gives each workload strong isolation from its neighbors. Now, why are the other options incorrect? Option A is incorrect. It shares the host operating system kernel rather than bringing its own full OS. Option B is incorrect. A cluster is a group of machines working together, not a single isolated compute unit. Option C is incorrect. A stand-alone server is just one physical machine and does not describe a virtualized, isolated unit. Therefore, D, Virtual machine, is the correct answer.
Play from 43:03 Question 31. A developer at a travel startup publishes a static website for visitors around the world. Which service would most help cut page load delay? Here are the options. Option A. CDN. Option B. VPC. Option C. API gateway. Option D. Application load balancer. Pause the video now if you want time to think about the answer.
The correct answer is A. CDN. A content delivery network keeps copies of the static pages at edge locations spread around the globe, so each visitor is served from a nearby point instead of one distant origin. That shortens the delay. A VPC only isolates a network, a load balancer spreads requests among servers in one place, and an API gateway fronts API calls. Now, why are the other options incorrect? Option B is incorrect. It provides a private network boundary and does nothing to bring content closer to visitors. Option C is incorrect. It manages and secures API calls, not delivery of static site files worldwide. Option D is incorrect. It spreads traffic among servers in a region and does not shorten the distance to far-away users. Therefore, A, CDN, is the correct answer.
Play from 44:35 Question 32. A startup is building a new web app and needs a relational database but wants as little administration work as possible. What should it pick? Here are the options. Option A. Installing and running a database engine on a virtual server. Option B. A tool for moving existing databases between environments. Option C. A cloud-hosted SQL database run as a managed service. Option D. A hybrid database arrangement spanning cloud and local systems. Pause the video now if you want time to think about the answer.
The correct answer is C. A cloud-hosted SQL database run as a managed service. With a managed relational service, the provider takes care of patching, backups, failover, and scaling, so the team simply uses the database. Running a database engine on a server leaves all that work with the team, a hybrid layout adds more moving parts, and a migration service only moves data from one place to another. Now, why are the other options incorrect? Option A is incorrect. The team would have to patch, back up, and maintain the server and the database software itself. Option B is incorrect. It is a tool for moving existing data, not a place to run a new application's database. Option D is incorrect. Spanning cloud and local systems adds complexity and more administration. Therefore, C, A cloud-hosted SQL database run as a managed service, is the correct answer.
Play from 46:23 Question 33. A hospital network must keep patient records private, tamper-proof, kept for at least five years, and backed up into a cloud storage bucket. Which approach satisfies all of this? Here are the options. Option A. Daily full backup. Option B. Data replication. Option C. Bulk recovery. Option D. Object Lock. Pause the video now if you want time to think about the answer.
The correct answer is D. Object Lock. Object Lock gives write-once, read-many protection on cloud bucket objects, so records cannot be changed or deleted until the retention period, such as five years, ends. Combined with access controls for privacy, that meets every item on the list. The other choices copy data but cannot enforce immutability or a retention period. Now, why are the other options incorrect? Option A is incorrect. It makes copies on a schedule, but nothing stops those copies from being altered or removed early. Option B is incorrect. It duplicates data to other places, including any unwanted changes, and enforces no retention. Option C is incorrect. It is about restoring lots of data after a failure, not about protecting records for compliance. Therefore, D, Object Lock, is the correct answer.
Play from 48:00 Question 34. A software vendor's highly available application currently runs on a local VM cluster, and the vendor now wants to serve outside users. Which migration type fits best? Here are the options. Option A. On-premises to cloud. Option B. Cloud to on-premises. Option C. On-premises to on-premises. Option D. Cloud to cloud. Pause the video now if you want time to think about the answer.
The correct answer is A. On-premises to cloud. The application starts in a local data center and needs to reach outside users, so the move is from on-premises into a cloud. Cloud providers offer wide geographic reach and redundant capacity that suit a highly available service for external audiences. Now, why are the other options incorrect? Option B is incorrect. That is the opposite direction and would not help reach outside users. Option C is incorrect. Staying in local data centers gives no easier way to reach external users. Option D is incorrect. That applies when the workload already lives in a cloud, which is not the starting point here. Therefore, A, On-premises to cloud, is the correct answer.
Play from 49:33 Question 35. A teammate asks you to define elasticity as a cloud characteristic. Which description is accurate? Here are the options. Option A. Fixed capacity. Option B. Data encryption. Option C. Resources that grow and shrink on their own as load changes. Option D. Backup frequency. Pause the video now if you want time to think about the answer.
The correct answer is C. Resources that grow and shrink on their own as load changes. Elasticity is the ability to add resources when load rises and release them when it falls, without manual work. Fixed capacity, encryption, and backup timing are different topics. Now, why are the other options incorrect? Option A is incorrect. A set amount that never changes is the opposite of elastic. Option B is incorrect. Encryption protects data confidentiality and has nothing to do with scaling. Option D is incorrect. How often backups run is a data protection setting, not a scaling behavior. Therefore, C, Resources that grow and shrink on their own as load changes, is the correct answer.
Play from 51:02 Question 36. A startup moving to the cloud wants its developers spending their time writing code, not looking after servers and operating systems. Which option fits best? Here are the options. Option A. Deploying IaaS. Option B. Subscribing to PaaS. Option C. Configuring FaaS. Option D. Purchasing SaaS. Pause the video now if you want time to think about the answer.
The correct answer is B. Subscribing to PaaS. With a platform service, the provider runs the servers, operating systems, runtimes, and tooling, leaving developers free to write and deploy code. IaaS hands the servers back to manage, FaaS covers only small event-driven functions, and SaaS is finished software that is not meant for custom development. Now, why are the other options incorrect? Option A is incorrect. Developers would still have to build and manage the virtual machines and the software stack. Option C is incorrect. It suits small event-triggered functions and is not a full platform for general applications. Option D is incorrect. It delivers ready-made software to use, not an environment to build custom code. Therefore, B, Subscribing to PaaS, is the correct answer.
Play from 52:41 Question 37. An operations lead needs to host a small service that is called rarely but must answer quickly when it is. Which compute option is the cheapest fit? Here are the options. Option A. Serverless. Option B. Container. Option C. Edge. Option D. Virtual machine. Pause the video now if you want time to think about the answer.
The correct answer is A. Serverless. Serverless runs code only when a request arrives and bills for that execution time, so a rarely used service costs almost nothing while idle, and it still scales and responds quickly. The other choices keep capacity reserved or running all the time and charge for it. Now, why are the other options incorrect? Option B is incorrect. Containers need a host and allocated resources that keep costing money between requests. Option C is incorrect. Edge deployments rely on dedicated distributed infrastructure that is paid for continuously. Option D is incorrect. A virtual machine is billed while it runs, whether or not anyone uses it. Therefore, A, Serverless, is the correct answer.
Play from 54:10 Question 38. After fixing a defect in an internal application, a developer wants the change saved with a record so the project history is preserved. What should they do? Here are the options. Option A. Code review. Option B. Code refactor. Option C. Code request. Option D. Code commit. Pause the video now if you want time to think about the answer.
The correct answer is D. Code commit. Committing saves the fix into version control together with a message, which adds it to the permanent project history. Reviews, refactors, and requests do not by themselves create that record. Now, why are the other options incorrect? Option A is incorrect. A review checks the work for quality, but it does not itself store the change in history. Option B is incorrect. Refactoring restructures existing code and is not how a fix gets recorded. Option C is incorrect. This is not a standard version control step. Therefore, D, Code commit, is the correct answer.
Play from 55:29 Question 39. Which statement best describes what makes a disaster recovery site a hot site? Here are the options. Option A. The standby servers share a cluster with the servers at the primary site. Option B. Every server is mirrored from the primary site and kept running online. Option C. Offline copies of the servers are refreshed from the primary site each hour. Option D. User traffic is split evenly across the primary and standby servers. Pause the video now if you want time to think about the answer.
The correct answer is B. Every server is mirrored from the primary site and kept running online. A hot site mirrors the production environment closely: every server is copied over from the main site and stays up and running, so failover can happen almost instantly. Warm and cold sites hold offline or partly prepared gear instead. Now, why are the other options incorrect? Option A is incorrect. Sharing a cluster with the main site is a local high availability design and not what defines a hot site. Option C is incorrect. Offline hourly copies describe a warm or cold approach. Option D is incorrect. Splitting traffic across both sites is an active-active design, not the defining trait of a hot site. Therefore, B, Every server is mirrored from the primary site and kept running online, is the correct answer.
Play from 57:16 Question 40. A research group runs short-term, non-critical jobs. Which purchasing model lowers cost the most? Here are the options. Option A. Spot instance. Option B. Reserved. Option C. Pay-as-you-go. Option D. Dedicated host. Pause the video now if you want time to think about the answer.
The correct answer is A. Spot instance. Spot capacity is spare provider capacity sold at steep discounts, and the provider may take it back, which is fine for brief, low-stakes jobs. Reserved pricing needs a long commitment, pay-as-you-go costs more per hour, and dedicated hosts are the priciest. Now, why are the other options incorrect? Option B is incorrect. It rewards a long commitment, which does not fit short-term work. Option C is incorrect. On-demand rates are higher than spare-capacity discounts. Option D is incorrect. A whole physical server is set aside for you, which is the most expensive route. Therefore, A, Spot instance, is the correct answer.
Play from 58:38 Question 41. A video streaming startup needs a delay-sensitive service to keep running even if a whole region goes dark, so it will be spread over several regions. Which concept fits best? Here are the options. Option A. Hardware passthrough. Option B. Stand-alone container. Option C. Clustering. Option D. Cloning. Pause the video now if you want time to think about the answer.
The correct answer is C. Clustering. Grouping several nodes that sit in different regions lets traffic shift automatically when one of them fails, so the service keeps answering quickly. The other choices either copy a resource, expose a device, or leave a single point of failure. Now, why are the other options incorrect? Option A is incorrect. Passthrough gives a virtual machine direct use of a physical device and does nothing for uptime. Option B is incorrect. A lone container is a single point of failure and cannot span regions by itself. Option D is incorrect. Cloning only duplicates a resource, with no failover or geographic spread. Therefore, C, Clustering, is the correct answer.
Play from 1:00:00 Question 42. A startup adopts a fully managed database offering from its cloud vendor. Which item must the startup's own team still plan for? Here are the options. Option A. Installing the service. Option B. Configuring IAM. Option C. Compiling sources. Option D. Applying OS updates. Pause the video now if you want time to think about the answer.
The correct answer is B. Configuring IAM. With a managed offering, the vendor takes care of the build, the server updates, and the setup. Deciding who may reach the service and what each person may do stays with the customer, so identity and access settings are the item to plan. Now, why are the other options incorrect? Option A is incorrect. The vendor sets up the service as part of the managed offer. Option C is incorrect. Building software from source is hidden behind the managed service. Option D is incorrect. The provider keeps the underlying operating system patched. Therefore, B, Configuring IAM, is the correct answer.
Play from 1:01:26 Question 43. A consultant is designing a cloud environment for a large bank. Which two requirements should rank highest? (Choose two.) Here are the options. Option A. Compliance. Option B. Cost. Option C. Network. Option D. Security. Option E. Compute. Option F. Performance. Option G. Storage. Pause the video now if you want time to think about the answer.
The correct answers are A. Compliance, and D. Security. Banks handle sensitive money data and answer to strict rules, so protecting that data and proving adherence to the rules come before everything else. Price, speed, and raw capacity still matter, but they follow after those two. Now, why are the other options incorrect? Option B is incorrect. Spend matters, yet it ranks below protection and rule-following for a bank. Option C is incorrect. Network design supports the top goals but is not itself the leading demand. Option E is incorrect. Processing power is a general infrastructure need, not specific to finance. Option F is incorrect. Speed is nice to have but gives way to regulation and data protection. Option G is incorrect. Capacity matters, but the controls around the stored data matter more. Therefore, A and D are the correct answers.
Play from 1:03:12 Question 44. A support team wants software that reads customer emails and works out whether the writers sound pleased or upset. Which AI/ML technology does this? Here are the options. Option A. Computer vision. Option B. Sentiment analysis. Option C. Visual recognition. Option D. Text recognition. Pause the video now if you want time to think about the answer.
The correct answer is B. Sentiment analysis. This technology reads written language and decides whether the mood is positive, negative, or neutral. That is exactly the job of judging how a message feels, which the other options cannot do. Now, why are the other options incorrect? Option A is incorrect. It interprets images and video, not the mood of written words. Option C is incorrect. It spots things in pictures rather than reading tone in text. Option D is incorrect. It pulls characters out of images, which is OCR, and says nothing about mood. Therefore, B, Sentiment analysis, is the correct answer.
Play from 1:04:36 Question 45. A SaaS firm compares hosting its platform in one region versus several. Which cost category would climb the most because data moves between locations? Here are the options. Option A. Power and cooling. Option B. Network. Option C. Storage. Option D. Compliance and regulation. Pause the video now if you want time to think about the answer.
The correct answer is B. Network. Traffic that crosses region borders is billed as outbound transfer, and a multi-region design generates plenty of it. A single region avoids most of those charges, so networking is where the bill grows fastest. Now, why are the other options incorrect? Option A is incorrect. These are provider facility costs that do not shift with the customer's layout. Option C is incorrect. Replicated copies add some storage cost, but transfer is the bigger jump. Option D is incorrect. These costs follow legal duties, not how many regions are used. Therefore, B, Network, is the correct answer.
Play from 1:05:58 Question 46. A startup wants full administrative control of the operating system on its cloud servers so it can install its own custom software stack. Which service model fits? Here are the options. Option A. IaaS. Option B. Software as a service. Option C. FaaS. Option D. PaaS. Pause the video now if you want time to think about the answer.
The correct answer is A. IaaS. With infrastructure as a service the provider supplies virtual hardware and the customer manages everything from the operating system upward. That gives the most control over the OS of any of the common service models. Now, why are the other options incorrect? Option B is incorrect. The provider runs the application and its operating system, so the customer has almost no OS control. Option C is incorrect. Function platforms hide the operating system entirely and only run small pieces of code. Option D is incorrect. The provider manages the operating system and runtime so developers just deploy code. Therefore, A, IaaS, is the correct answer.
Play from 1:07:27 Question 47. A provider launches a new serverless platform that meets every security regulation. A firm runs its code there, and attackers get in because the firm's access credentials were exposed. Who bears the responsibility? Here are the options. Option A. Hacker. Option B. Code repository. Option C. Customer. Option D. Cloud service provider. Pause the video now if you want time to think about the answer.
The correct answer is C. Customer. Under shared responsibility, the provider secures the platform, but secrets and credentials belong to whoever writes and deploys the code. The platform was compliant and the leak came from how the firm handled its own keys, so the fault sits with the firm. Now, why are the other options incorrect? Option A is incorrect. The attacker took advantage of the leak but did not cause it. Option B is incorrect. A repository is only a tool, and the customer decides what is stored in it. Option D is incorrect. The platform was compliant, and the weakness was in the customer's handling of credentials. Therefore, C, Customer, is the correct answer.
Play from 1:08:56 Question 48. A company wants to encrypt everything it stores with a public cloud using keys it holds itself, but the provider refuses to allow customer-held keys. Which principle is the provider breaking? Here are the options. Option A. Data retention. Option B. Data ownership. Option C. Data classification. Option D. Data sovereignty. Pause the video now if you want time to think about the answer.
The correct answer is B. Data ownership. Owning your data means deciding how it is protected, and that includes locking it with your own keys. Blocking those keys takes that control away from the customer. Now, why are the other options incorrect? Option A is incorrect. Retention is about how long data is kept, not who controls the keys. Option C is incorrect. Classification is about labeling data by sensitivity. Option D is incorrect. Sovereignty deals with which country's laws apply to the data's location. Therefore, B, Data ownership, is the correct answer.
Play from 1:10:21 Question 49. After updating an app, a developer must pick a release approach that is quick and easy and can run on a pair of matching environments. Which should be chosen? Here are the options. Option A. Canary. Option B. Rolling. Option C. In-place. Option D. Blue-green. Pause the video now if you want time to think about the answer.
The correct answer is D. Blue-green. With two matching environments, one live and one idle, you push the new version to the idle side and then flip traffic over. That is simple and fast, and flipping back is just as easy if something breaks. Now, why are the other options incorrect? Option A is incorrect. It exposes the change to a small slice of users first, which is slower and more involved. Option B is incorrect. It swaps instances in batches over time, so it is not the quick option. Option C is incorrect. It overwrites the running environment, adding risk and ignoring the second platform. Therefore, D, Blue-green, is the correct answer.
Play from 1:11:45 Question 50. A data team runs batch jobs in containers and needs the results saved as files they can open later. What should they use? Here are the options. Option A. Ephemeral storage. Option B. Persistent volumes. Option C. Optical disk mount. Option D. Standard output. Pause the video now if you want time to think about the answer.
The correct answer is B. Persistent volumes. A persistent volume lives apart from the container, so files written there survive restarts and even deletion of the container. That makes it the right place for results you want to read afterward. Now, why are the other options incorrect? Option A is incorrect. It disappears once the container stops, so the results would be gone. Option C is incorrect. This is not a practical way to store container data in the cloud. Option D is incorrect. It is a text stream for the console or logs, not a place to keep files. Therefore, B, Persistent volumes, is the correct answer.
Play from 1:13:01 Question 51. A retailer wants its web tier to stay online even when one availability zone has an outage. Which approach achieves that? Here are the options. Option A. Disabling load balancing. Option B. Vertical scaling. Option C. Spreading instances across several zones. Option D. Relying on one big server. Pause the video now if you want time to think about the answer.
The correct answer is C. Spreading instances across several zones. When copies of the workload run in separate zones, the surviving zones keep serving users after one zone fails. Now, why are the other options incorrect? Option A is incorrect. Without load balancing, traffic cannot be steered away from a failed zone. Option B is incorrect. It only makes a single instance larger and leaves it in one zone. Option D is incorrect. One server is a single point of failure. Therefore, C, Spreading instances across several zones, is the correct answer.
Play from 1:14:17 Question 52. To prepare for a regulatory certification, a sysadmin must change the company's logging policy so records are still on hand throughout the audit window. What should be put in place? Here are the options. Option A. Log tracing. Option B. Log collection. Option C. Log retention. Option D. Log aggregation. Pause the video now if you want time to think about the answer.
The correct answer is C. Log retention. The need here is about how long records are kept, so they can still be produced when auditors ask. Setting a keep-for-this-long rule on the logs covers exactly that. Now, why are the other options incorrect? Option A is incorrect. It follows a request across services and has nothing to do with storage duration. Option B is incorrect. It gathers logs from systems but does not decide how long they are kept. Option D is incorrect. It combines logs in one place without setting any keeping period. Therefore, C, Log retention, is the correct answer.
Play from 1:15:40 Question 53. A developer turns on API access for a service account in a SaaS product so a script can call its REST interface. Which two items are needed to authenticate? (Choose two.) Here are the options. Option A. Username. Option B. Secret key. Option C. Cookie. Option D. Workstation name. Option E. Client ID. Option F. Password. Pause the video now if you want time to think about the answer.
The correct answers are B. Secret key, and E. Client ID. Automation against a REST interface usually relies on an application identity pair: one value says who is calling and a private value proves it. Human login details and browser artifacts are not how scripts normally sign in. Now, why are the other options incorrect? Option A is incorrect. This is a human login detail, not an application credential. Option C is incorrect. Cookies belong to browser sessions, not scripted calls. Option D is incorrect. The name of a computer plays no part in API sign-in. Option F is incorrect. Passwords are for people, not for programmatic access. Therefore, B and E are the correct answers.
Play from 1:17:16 Question 54. A developer runs a web app inside a container and needs outside clients to be able to reach it. Which feature makes that possible? Here are the options. Option A. Host affinity. Option B. Port mapping. Option C. Forward proxy. Option D. Network orchestration. Pause the video now if you want time to think about the answer.
The correct answer is B. Port mapping. Publishing a container's port on the host lets outside traffic arrive at the application running inside. Without that link, the service stays unreachable from outside. Now, why are the other options incorrect? Option A is incorrect. It controls where workloads are placed, not how they are reached. Option C is incorrect. It relays outbound client requests and does not publish a service. Option D is incorrect. It automates network setup in general but does not itself expose a container service. Therefore, B, Port mapping, is the correct answer.
Play from 1:18:35 Question 55. Two companies sign a binding document in which both promise to encrypt any data passed between their environments. What kind of agreement is this? Here are the options. Option A. Subscription. Option B. Contractual. Option C. Litigation. Option D. Regulatory. Pause the video now if you want time to think about the answer.
The correct answer is B. Contractual. A signed deal between two parties that sets out duties both must follow is a contract. Here the duty is encrypting shared data, and it binds only the two signers. Now, why are the other options incorrect? Option A is incorrect. That describes a recurring service plan or payment arrangement. Option C is incorrect. That refers to lawsuits and legal disputes, not a signed agreement. Option D is incorrect. Those rules come from governments or industry bodies, not from a deal between two firms. Therefore, B, Contractual, is the correct answer.
Play from 1:19:53 Question 56. A hospital IT manager wants to know how much information could be lost at most after an outage. Which item drives that figure the most? Here are the options. Option A. Average time needed to recover. Option B. Recovery point objective. Option C. Recovery time objective. Option D. The overall BCDR program. Pause the video now if you want time to think about the answer.
The correct answer is B. Recovery point objective. This target is measured as a span of time back from the failure and says how far the data may roll back. That span sets the most data you can afford to lose, which guides how often backups or replication must run. Now, why are the other options incorrect? Option A is incorrect. It measures how long a repair takes, not how much data disappears. Option C is incorrect. It sets how long downtime may last, not the data loss. Option D is incorrect. It is the umbrella plan that contains the targets rather than the data-loss measure itself. Therefore, B, Recovery point objective, is the correct answer.
Play from 1:21:18 Question 57. A media company must store many kinds of unstructured content whose shape keeps changing. Which database type suits it best? Here are the options. Option A. Graph. Option B. Vector. Option C. Non-relational. Option D. Relational. Pause the video now if you want time to think about the answer.
The correct answer is C. Non-relational. Databases without a fixed table design accept varied formats and let the structure evolve, which suits loose content that keeps changing. Now, why are the other options incorrect? Option A is incorrect. It focuses on links between items, not general loose content. Option B is incorrect. It targets embeddings used for similarity search. Option D is incorrect. It demands a fixed schema in tables, which fights against constant change. Therefore, C, Non-relational, is the correct answer.
Play from 1:22:34 Question 58. An architect wants a cloud-native application design that is easiest to maintain and whose parts are decentralized and loosely coupled. Which design fits? Here are the options. Option A. Mainframe. Option B. Microservices. Option C. Monolithic. Option D. Hybrid cloud. Pause the video now if you want time to think about the answer.
The correct answer is B. Microservices. Splitting an application into many small services, each deployed and updated on its own and talking through APIs, gives loose coupling and easy upkeep. Now, why are the other options incorrect? Option A is incorrect. It is a legacy platform and not a modern cloud-native design. Option C is incorrect. Everything sits in one tightly bound codebase, the opposite of decoupled. Option D is incorrect. It describes mixing on-premises and cloud hosting, not an application design. Therefore, B, Microservices, is the correct answer.
Play from 1:23:52 Question 59. A storage admin is choosing drives. Which connection interface is typically associated with solid-state drives rather than spinning disks? Here are the options. Option A. SATA. Option B. iSCSI. Option C. NVMe. Option D. Serial-attached SCSI. Pause the video now if you want time to think about the answer.
The correct answer is C. NVMe. This interface was built for flash storage and talks straight over PCIe, taking advantage of very low delay. In practice it is found on solid-state drives, while the rest are used with both drive types. Now, why are the other options incorrect? Option A is incorrect. It is common on both solid-state and spinning drives. Option B is incorrect. It carries storage traffic over a network and can sit in front of either drive type. Option D is incorrect. Enterprise systems use it with both kinds of drive. Therefore, C, NVMe, is the correct answer.
Play from 1:25:12 Question 60. Which area of computer science teaches machines to recognize and interpret objects and people appearing in pictures and video? Here are the options. Option A. Natural language processing. Option B. Computer vision. Option C. Facial recognition. Option D. Image reconstruction. Pause the video now if you want time to think about the answer.
The correct answer is B. Computer vision. This field covers anything a machine does to make sense of visual input, from spotting objects to following people on screen. Now, why are the other options incorrect? Option A is incorrect. It deals with written and spoken language, not visuals. Option C is incorrect. It is just one narrow use of the wider field. Option D is incorrect. It is a technique for rebuilding images from data, a smaller piece of the field. Therefore, B, Computer vision, is the correct answer.
Play from 1:26:26 Question 61. A storage architect is sizing volumes for a new database server. Which piece of information is required to compute the IOPS the storage must provide? Here are the options. Option A. Latency. Option B. Memory utilization. Option C. Block size. Option D. Network capacity. Pause the video now if you want time to think about the answer.
The correct answer is C. Block size. An IOPS figure only turns into real data movement once you know how large each operation is. Small blocks mean a great many operations to move a given amount of data, while large blocks need far fewer. That is why the sizing math cannot be done without knowing the size of each block. Now, why are the other options incorrect? Option A is incorrect. Latency is an outcome of how the storage performs, not an input to the IOPS sizing math. Option B is incorrect. Memory use has no bearing on how many storage operations a workload needs. Option D is incorrect. Network capacity relates to bandwidth, not to the count of storage operations per second. Therefore, C, Block size, is the correct answer.
Play from 1:27:56 Question 62. A state agency in the public sector plans to move its workloads off its own servers and onto a cloud platform. Which two factors should weigh most heavily in that decision? (Choose two.) Here are the options. Option A. Service availability. Option B. Regulatory. Option C. Implementation timeline. Option D. Compliance. Option E. Firewall capabilities. Option F. IaaS vs. SaaS. Pause the video now if you want time to think about the answer.
The correct answers are B. Regulatory, and D. Compliance. Public sector bodies answer to strict legal frameworks, so any cloud they pick has to satisfy mandated rules and standards such as FedRAMP. Those obligations come first and then shape every other choice in the move, which is why regulatory and compliance requirements top the list. Now, why are the other options incorrect? Option A is incorrect. Availability matters, but it has to be delivered within the legal limits, so it ranks below the mandated rules. Option C is incorrect. A schedule is a project planning detail that gets set once the mandatory requirements are known. Option E is incorrect. Firewall features are a design detail that follows from the rules the agency must obey. Option F is incorrect. Picking a service model is a later design decision, not the top concern for a regulated agency. Therefore, B and D are the correct answers.
Play from 1:29:46 Question 63. A developer at a startup is deploying a cloud app that is built around web sockets. What kind of application is most likely being deployed? Here are the options. Option A. Chat. Option B. File transfer. Option C. Data visualization. Option D. Image-sharing. Pause the video now if you want time to think about the answer.
The correct answer is A. Chat. A web socket keeps one connection open so the client and server can send data to each other at any moment. Messaging tools rely on exactly that behavior, because new messages must show up instantly without the client asking over and over. Now, why are the other options incorrect? Option B is incorrect. Moving files works fine over ordinary HTTP requests and does not need a constantly open two-way channel. Option C is incorrect. Dashboards usually refresh on a timer through standard API calls rather than needing a permanent socket. Option D is incorrect. Sharing pictures is typically handled with regular HTTP uploads and downloads. Therefore, A, Chat, is the correct answer.
Play from 1:31:16 Question 64. Before checking its cloud environment for weaknesses, a cloud engineer plans the work. What should the engineer do first? Here are the options. Option A. Research the CVEs. Option B. Define the scope. Option C. Remediate the vulnerabilities. Option D. Assess the systems. Pause the video now if you want time to think about the answer.
The correct answer is B. Define the scope. Every vulnerability program begins by deciding exactly which assets and environments are in play. Until that boundary is set, there is nothing to scan, rank, or fix, so everything else depends on this opening move. Now, why are the other options incorrect? Option A is incorrect. Looking up CVEs makes sense only after scan findings exist to look up. Option C is incorrect. Fixing problems comes at the end, once they have been found and ranked. Option D is incorrect. Assessing systems needs a defined target list, so it follows the scoping step. Therefore, B, Define the scope, is the correct answer.
Play from 1:32:37 Question 65. After a scan completes, a cloud developer reviews the findings and documents the confirmed true positives. Which vulnerability management step does this describe? Here are the options. Option A. Remediation. Option B. Scanning scope. Option C. Assessment. Option D. Identification. Pause the video now if you want time to think about the answer.
The correct answer is C. Assessment. This is the stage where raw scanner output gets examined, false alarms are separated from genuine issues, and the real ones are written down. Evaluating and validating results like this is the heart of the assessment phase, which sits between the scan itself and the fixes. Now, why are the other options incorrect? Option A is incorrect. Remediation is about applying fixes, which happens after the findings are reviewed. Option B is incorrect. Scope is decided up front to choose what gets scanned, not to review results. Option D is incorrect. Identification is the discovery of possible issues, not the checking of findings for accuracy. Therefore, C, Assessment, is the correct answer.
Play from 1:34:10 Question 66. A streaming company's website slows down every month when new episodes go live, and those release dates are known well in advance. Which scaling approach lets the company add capacity based on that forecast? Here are the options. Option A. Load. Option B. Scheduled. Option C. Manual. Option D. Event. Pause the video now if you want time to think about the answer.
The correct answer is B. Scheduled. Because the surges follow a predictable calendar, capacity can be added ahead of time on a timetable. Resources are already in place when viewers arrive, and nobody has to step in during the rush. Now, why are the other options incorrect? Option A is incorrect. Load-based scaling reacts only after traffic has already climbed, so users still feel the slowdown first. Option C is incorrect. Manual scaling depends on a person acting at the right moment and uses none of the known timing. Option D is incorrect. Event-driven scaling waits for a trigger to fire instead of using the forecast release dates. Therefore, B, Scheduled, is the correct answer.
Play from 1:35:35 Question 67. A retailer runs its own server room and has moved a few of its workloads to a cloud provider, while the rest stay in the server room. Which deployment model describes this setup best? Here are the options. Option A. Hybrid. Option B. Community. Option C. Public. Option D. Private. Pause the video now if you want time to think about the answer.
The correct answer is A. Hybrid. Mixing local infrastructure with services in a provider's cloud is the textbook definition of a hybrid design. Some workloads live in the company's own facility and others run remotely, and the two are used together. Now, why are the other options incorrect? Option B is incorrect. A community cloud is shared by several groups with common needs, which is not the case here. Option C is incorrect. A public cloud means everything runs on the provider's shared platform, with nothing kept in-house. Option D is incorrect. A private cloud is dedicated to one organization and does not split workloads with an outside provider. Therefore, A, Hybrid, is the correct answer.
Play from 1:37:02 Question 68. Which of these is the standard public listing of known security vulnerabilities? Here are the options. Option A. Zero Trust. Option B. CWE, the list of software weakness categories. Option C. Data Security Standards. Option D. CVE, the public catalog of disclosed vulnerabilities. Pause the video now if you want time to think about the answer.
The correct answer is D. CVE, the public catalog of disclosed vulnerabilities. CVE entries each describe one specific flaw that has been publicly disclosed and given its own tracking number. Security teams use that shared catalog to look up and talk about known problems in the same way. Now, why are the other options incorrect? Option A is incorrect. Zero Trust is a design philosophy for access control, not a list of flaws. Option B is incorrect. CWE groups types of coding weaknesses in general and does not track individual known vulnerabilities. Option C is incorrect. Data security standards are compliance rules such as PCI DSS, not a vulnerability list. Therefore, D, CVE, the public catalog of disclosed vulnerabilities, is the correct answer.
Play from 1:38:43 Question 69. A media company wants to move its on-premises archive of unstructured data into the cloud, and the solution must be low in cost and able to scale enormously. Which storage type fits best? Here are the options. Option A. SAN. Option B. Object. Option C. File. Option D. Block. Pause the video now if you want time to think about the answer.
The correct answer is B. Object. Data is kept as self-describing objects in a flat layout, which lets capacity grow almost without limit at a low price per gigabyte. That combination is why cloud teams default to it for large pools of unstructured content. Now, why are the other options incorrect? Option A is incorrect. A SAN is a costly enterprise storage network and a poor match for a low-cost cloud move. Option C is incorrect. File storage relies on a folder hierarchy that scales less well and costs more for huge unstructured sets. Option D is incorrect. Block storage is meant for individual volumes, and its price per unit is higher for bulk content. Therefore, B, Object, is the correct answer.
Play from 1:40:15 Question 70. A systems administrator edits the iptables rules so that traffic to and from the address 104.225.110.203 is dropped in both directions. Which vulnerability management step does this represent? Here are the options. Option A. Identification. Option B. Assessment. Option C. Remediation. Option D. Scanning scope. Pause the video now if you want time to think about the answer.
The correct answer is C. Remediation. Cutting off a hostile address with a firewall rule is a hands-on action that reduces or removes the risk. Taking a protective step like this is what remediation means, as opposed to finding, measuring, or planning what to scan. Now, why are the other options incorrect? Option A is incorrect. Identification is the discovery of a problem, which would have happened before any blocking rule was written. Option B is incorrect. Assessment ranks and analyzes findings rather than putting a technical control in place. Option D is incorrect. Scope only defines which assets get scanned and does not involve filtering any traffic. Therefore, C, Remediation, is the correct answer.
Play from 1:41:48 Question 71. A backup engineer cuts how long the backup jobs take to run, and the CFO now wants proof that restores will return everything with nothing lost. Which backup feature should the engineer test? Here are the options. Option A. Integrity. Option B. Schedule. Option C. Encryption. Option D. Retention. Pause the video now if you want time to think about the answer.
The correct answer is A. Integrity. Checking integrity confirms that the saved data is complete and uncorrupted, and that a restore will bring it all back. Since the concern is a faster job possibly losing something, proving the copies are intact is the right test. Now, why are the other options incorrect? Option B is incorrect. The schedule only sets when jobs run and says nothing about whether a restore is complete. Option C is incorrect. Encryption guards confidentiality but does not prove the data comes back whole. Option D is incorrect. Retention controls how long copies are kept, not whether they can be restored. Therefore, A, Integrity, is the correct answer.
Play from 1:43:15 Question 72. Which command lets a developer upload local commits to the shared remote repository? Here are the options. Option A. git commit. Option B. git pull. Option C. git push. Option D. git fetch. Pause the video now if you want time to think about the answer.
The correct answer is C. git push. Pushing sends the commits you made on your machine up to the server copy, so teammates can see them. Every other choice either stays local or brings data down instead of sending it up. Now, why are the other options incorrect? Option A is incorrect. A commit records changes in the local repository only and never contacts the server. Option B is incorrect. A pull downloads remote changes and merges them into the local branch, the opposite direction. Option D is incorrect. A fetch only downloads remote updates and sends nothing back. Therefore, C, git push, is the correct answer.
Play from 1:44:32 Question 73. A platform team maintains infrastructure templates for its cloud deployments. Which practice lowers the likelihood of a misconfiguration slipping in? Here are the options. Option A. Pushing updates directly to the main branch. Option B. Editing the files through the repository web page. Option C. Running a fetch right after each commit. Option D. Submitting changes through pull requests for review. Pause the video now if you want time to think about the answer.
The correct answer is D. Submitting changes through pull requests for review. A pull request puts a change in front of other people before it is merged. A second pair of eyes tends to catch mistakes in a template early, long before the faulty setting could reach a live environment. Now, why are the other options incorrect? Option A is incorrect. Writing straight to main skips any review, so a bad setting can go live right away. Option B is incorrect. Editing in the browser sidesteps the normal review and testing workflow and raises the chance of errors. Option C is incorrect. A fetch merely downloads updates and adds no check on the quality of a change. Therefore, D, Submitting changes through pull requests for review, is the correct answer.
Play from 1:46:05 Question 74. A security analyst at a fintech firm worries about flaws in container images pulled from public registries. What should the analyst recommend? Here are the options. Option A. Using watermarked images. Option B. Choosing images that include a built-in application firewall. Option C. Relying on cryptographically signed images. Option D. Using CIS-hardened images. Pause the video now if you want time to think about the answer.
The correct answer is D. Using CIS-hardened images. Images built to Center for Internet Security benchmarks start out with unneeded components removed, patches applied, and safer defaults set. That shrinks the number of flaws inside the image itself, which is exactly the analyst's worry. Now, why are the other options incorrect? Option A is incorrect. A watermark marks ownership and does nothing about weaknesses in the image contents. Option B is incorrect. A firewall guards a running container from outside traffic, but the flaws inside the image remain. Option C is incorrect. A signature proves who published the image and that it was not altered, yet it does not remove any vulnerabilities. Therefore, D, Using CIS-hardened images, is the correct answer.
Play from 1:47:45 Question 75. An insurer must keep every customer policy for over a decade to satisfy regulators. Which cloud storage tier is the cheapest way to do this? Here are the options. Option A. Cold. Option B. Warm. Option C. Archive. Option D. Hot. Pause the video now if you want time to think about the answer.
The correct answer is C. Archive. Data that must be kept for years but is rarely read belongs in the tier with the lowest storage price. The trade-off is slower and costlier retrieval, which is acceptable for records that sit untouched for long periods. Now, why are the other options incorrect? Option A is incorrect. Cold storage is cheaper than hot, but it still costs more per gigabyte than the archive tier. Option B is incorrect. Warm sits in the middle on price and speed, so it wastes money on data that is rarely touched. Option D is incorrect. Hot storage gives the fastest access at the highest price, which is unnecessary here. Therefore, C, Archive, is the correct answer.
Play from 1:49:07 Question 76. A container is restarted and everything that was written inside it disappears. What kind of storage was the container using? Here are the options. Option A. Block. Option B. Object. Option C. Persistent volume. Option D. Ephemeral. Pause the video now if you want time to think about the answer.
The correct answer is D. Ephemeral. Ephemeral storage lives and dies with the container, so a restart wipes whatever was written there. To keep data across restarts it must be placed on storage that exists independently of the container. Now, why are the other options incorrect? Option A is incorrect. Block storage is a durable disk-style volume that survives restarts when attached. Option B is incorrect. Object storage is a durable service whose contents do not depend on a container's lifecycle. Option C is incorrect. A persistent volume exists precisely to keep data after a container restarts. Therefore, D, Ephemeral, is the correct answer.
Play from 1:50:28 Question 77. A software company is restructuring its application into microservices. Which description fits the new design? Here are the options. Option A. Loosely coupled architecture. Option B. Private cloud-managed services. Option C. Tightly coupled architecture. Option D. Public cloud-managed services. Pause the video now if you want time to think about the answer.
The correct answer is A. Loosely coupled architecture. Microservices split an application into small, independent pieces that each handle one job and talk through defined APIs. Because the pieces do not depend heavily on each other, teams can build, release, and scale them separately. Now, why are the other options incorrect? Option B is incorrect. Microservices can run on a private cloud, but that describes where they run, not the design pattern. Option C is incorrect. Tight coupling is the reverse of this approach and fits a single monolithic application. Option D is incorrect. A public cloud is a hosting choice, while microservices describe how the software is structured. Therefore, A, Loosely coupled architecture, is the correct answer.
Play from 1:51:59 Question 78. Ten engineers at a hosting firm each carry out cloud deployments by hand, and the results keep coming out different. Which approach best fixes this? Here are the options. Option A. Change ticketing. Option B. Configuration as code. Option C. Service logging. Option D. Deployment documentation. Pause the video now if you want time to think about the answer.
The correct answer is B. Configuration as code. Writing the setup down as code means every deployment runs from the same definition, which can be versioned and reviewed. Nobody has to improvise, so the engineers get the same result each time and the drift between deployments goes away. Now, why are the other options incorrect? Option A is incorrect. Tickets record that a change happened but do nothing to make deployments identical. Option C is incorrect. Logs show what occurred after the fact and cannot prevent differences while deploying. Option D is incorrect. Written instructions still depend on each person's interpretation, which is the cause of the problem. Therefore, B, Configuration as code, is the correct answer.
Play from 1:53:30 Question 79. A consultant recommends moving an accounting application to the cloud, using infrastructure that is shared among tenants to keep costs low. Which cloud model best matches? Here are the options. Option A. Hybrid. Option B. Public. Option C. Private. Option D. Community. Pause the video now if you want time to think about the answer.
The correct answer is B. Public. In a public cloud, many customers run on the same pooled hardware, so the provider spreads the expense across all of them. That sharing is what drives the lowest prices of any model. Now, why are the other options incorrect? Option A is incorrect. Hybrid blends several models and does not by itself deliver the cost savings of broad sharing. Option C is incorrect. A private cloud is dedicated to one organization, so no cost is split with others. Option D is incorrect. A community cloud is shared only within a small group, which means fewer tenants and a higher cost per member. Therefore, B, Public, is the correct answer.
Play from 1:54:52 Question 80. A security engineer finds that a service account's secrets were checked into the team's source repository. Which two actions should come first? (Choose two.) Here are the options. Option A. Scan the open ports. Option B. Rotate the credentials. Option C. Alter the cipher suite in use. Option D. Enable multifactor authentication. Option E. Disable the account. Option F. Stand up a VPN. Pause the video now if you want time to think about the answer.
The correct answers are B. Rotate the credentials, and E. Disable the account. Once secrets are visible in a repository, anyone could use them. The urgent move is to shut the account off so it cannot be abused, and then replace the secrets with new ones so the leaked copies are worthless. Now, why are the other options incorrect? Option A is incorrect. A port scan looks for open services and does nothing about leaked secrets. Option C is incorrect. Cipher settings concern encryption of traffic, not credentials that are already exposed. Option D is incorrect. Adding multifactor authentication does not undo the exposure of the secrets that already leaked. Option F is incorrect. A VPN does not resolve the underlying problem of exposed service account credentials. Therefore, B and E are the correct answers.
Play from 1:56:35 Question 81. A platform team at a retailer wants one tool to roll out the same template-defined, container-based services in its own data center and in a public cloud. Which tool fits best? Here are the options. Option A. Git. Option B. Jenkins. Option C. Terraform. Option D. Docker. Pause the video now if you want time to think about the answer.
The correct answer is C. Terraform. The clues are template-driven deployment and a hybrid target, which points to infrastructure as code. Terraform reads reusable definitions and talks to many providers, so one workflow can build out both a private data center and a public cloud. The other tools handle different parts of the software lifecycle. Now, why are the other options incorrect? Option A is incorrect. Git only tracks changes to files and does not deploy anything. Option B is incorrect. Jenkins runs build and release pipelines, but it does not provision infrastructure from templates. Option D is incorrect. Docker packages an app into a container, yet it has no built-in way to provision environments across providers. Therefore, C, Terraform, is the correct answer.
Play from 1:58:08 Question 82. A software startup wants to stay free of vendor lock-in while it evaluates cloud providers. Which two factors should it weigh before signing with one? (Choose two.) Here are the options. Option A. Service availability. Option B. Cost factor. Option C. Open-source technologies. Option D. Regulatory compliance. Option E. Storage options. Option F. Data portability. Pause the video now if you want time to think about the answer.
The correct answers are C. Open-source technologies, and F. Data portability. Lock-in happens when your tools and your data are tied to a single provider. Building on open standards keeps your choices open, and being able to move your data elsewhere with little effort means you can leave if you need to. The remaining factors matter when picking a vendor, but they do not protect you from getting stuck. Now, why are the other options incorrect? Option A is incorrect. Uptime is a quality measure and says nothing about how easily you can leave. Option B is incorrect. Price is a business concern, not a guard against being tied to one vendor. Option D is incorrect. Compliance is a requirement to meet, but it applies no matter which vendor you pick. Option E is incorrect. Storage features describe what a vendor offers, not how portable your workloads are. Therefore, C and F are the correct answers.
Play from 1:59:57 Question 83. After a ransomware incident at a logistics firm, the legal counsel asks the cloud admin to preserve the data from the impacted virtual machines for at least twelve months. Which concept does this describe? Here are the options. Option A. Integrity. Option B. Encryption. Option C. Recoverability. Option D. Retention. Pause the video now if you want time to think about the answer.
The correct answer is D. Retention. When a legal team says to keep data for a set length of time, that is a rule about how long records must be preserved. That is what retention means, and it applies regardless of how the incident itself was handled. Now, why are the other options incorrect? Option A is incorrect. Integrity is about data not being altered improperly, not about how long it is kept. Option B is incorrect. Encryption protects confidentiality and does not set a storage period. Option C is incorrect. Recoverability is the ability to restore systems or data, not a required keeping time. Therefore, D, Retention, is the correct answer.
Play from 2:01:21 Question 84. An online retailer wants shoppers to have a conversational, interactive way to resolve everyday problems on its site. Which technology would most likely supply this? Here are the options. Option A. Text recognition. Option B. Help page. Option C. Generative AI. Option D. Search engine. Pause the video now if you want time to think about the answer.
The correct answer is C. Generative AI. A conversational assistant that understands plain-language questions and replies on the spot gives visitors an interactive way to fix problems. Generative AI is the technology that powers that kind of experience. The other choices only show fixed content or convert images to text. Now, why are the other options incorrect? Option A is incorrect. Text recognition pulls characters out of images and does not hold a conversation. Option B is incorrect. A help page is static content that visitors read, not an interactive exchange. Option D is incorrect. A search engine returns links or documents rather than conversing to solve a problem. Therefore, C, Generative AI, is the correct answer.
Play from 2:02:51 Question 85. What do software teams rely on to record and follow the modifications made to source code over the life of a project? Here are the options. Option A. Code testing. Option B. Code versioning. Option C. Code drifting. Option D. Code control. Pause the video now if you want time to think about the answer.
The correct answer is B. Code versioning. Tracking every edit to source code, with the ability to browse history and roll back, is called versioning, and tools such as Git are built for exactly that. The other terms either describe unrelated activities or are not standard practices. Now, why are the other options incorrect? Option A is incorrect. Testing checks that code works, it does not record who changed what. Option C is incorrect. Drift describes configurations slowly diverging from their intended state, not a way to log changes. Option D is incorrect. This is not a recognized practice for keeping a history of changes. Therefore, B, Code versioning, is the correct answer.
Play from 2:04:16 Question 86. A finance team needs to run one short script at a fixed time every night. Which compute option is the most efficient choice? Here are the options. Option A. Virtual machine. Option B. Serverless function. Option C. Bare-metal server. Option D. Managed container. Pause the video now if you want time to think about the answer.
The correct answer is B. Serverless function. A job that runs one short script on a timer does not need a machine running all day. A serverless function starts only when triggered, bills for the brief time it runs, and needs no server upkeep. The other options leave you paying for and maintaining resources that sit idle most of the time. Now, why are the other options incorrect? Option A is incorrect. A VM must be kept running and patched even when the script is idle. Option C is incorrect. Dedicating physical hardware to one small script is costly and wasteful. Option D is incorrect. A container platform adds orchestration overhead that a single scheduled script does not need. Therefore, B, Serverless function, is the correct answer.
Play from 2:05:44 Question 87. A sales organization wants to move its customer relationship management software to the cloud and keep the IT workload as light as possible. Which service model is the best fit? Here are the options. Option A. IaaS. Option B. FaaS. Option C. SaaS. Option D. PaaS. Pause the video now if you want time to think about the answer.
The correct answer is C. SaaS. If the goal is minimal administration, the provider should run as much of the stack as possible. With software as a service, the vendor handles servers, the platform, updates and security, and the company simply uses the CRM. The other models leave the company managing at least the application itself, or are meant for small event-driven functions. Now, why are the other options incorrect? Option A is incorrect. IaaS leaves the customer managing operating systems and applications on rented infrastructure. Option B is incorrect. FaaS runs short event-driven code, not a full business application. Option D is incorrect. PaaS still requires the company to build and manage its own application. Therefore, C, SaaS, is the correct answer.
Play from 2:07:14 Question 88. A security engineer wants alerts when cloud servers are quietly hijacked to mine digital currency without permission. Which metric would most likely reveal this? Here are the options. Option A. Disk I/O. Option B. Processor usage percentage. Option C. Average memory utilization. Option D. Network packets. Pause the video now if you want time to think about the answer.
The correct answer is B. Processor usage percentage. Mining digital coins is computation heavy, so hijacked instances show processor usage pinned high for long stretches with no matching business workload. Watching that percentage is the simplest way to catch it. Other metrics may move too, but they are not the telltale sign. Now, why are the other options incorrect? Option A is incorrect. Mining is not storage heavy, so disk activity does not stand out. Option C is incorrect. Memory use stays fairly modest during mining and is not a reliable signal. Option D is incorrect. Mining traffic is light, so packet counts do not reveal it clearly. Therefore, B, Processor usage percentage, is the correct answer.
Play from 2:08:44 Question 89. A systems engineer is creating a standard corporate server image that starts from a publicly available one. What should be done to make that image secure? Here are the options. Option A. ACLs. Option B. Least privilege. Option C. Vulnerability scanning. Option D. Hardening. Pause the video now if you want time to think about the answer.
The correct answer is D. Hardening. Starting from a public image means you inherit unknown settings, extra services and possible weaknesses. Hardening addresses this by turning off what is not needed, closing ports, patching and tightening configuration before the image becomes the company standard. The other choices are useful controls but do not by themselves make the image itself secure. Now, why are the other options incorrect? Option A is incorrect. Access control lists limit who can reach resources, but they do not tighten the image contents. Option B is incorrect. Least privilege governs permissions granted to accounts, not the baseline build of the image. Option C is incorrect. Scanning only finds problems and does not remove or reconfigure anything. Therefore, D, Hardening, is the correct answer.
Play from 2:10:15 Question 90. An engineer designing sensor connectivity for a smart-building project needs a protocol that keeps message size and traffic small. Which one is commonly chosen for IoT? Here are the options. Option A. NFS. Option B. HTTPS. Option C. SMB. Option D. MQTT. Pause the video now if you want time to think about the answer.
The correct answer is D. MQTT. Small devices often run on weak links and limited power, so the protocol needs a tiny header and a publish and subscribe design. MQTT is built for that, which keeps the amount of data sent low. File sharing protocols and general web traffic carry far more overhead. Now, why are the other options incorrect? Option A is incorrect. NFS shares files over a network and is not designed for lightweight device messaging. Option B is incorrect. HTTPS is a general web protocol with heavier overhead per message. Option C is incorrect. SMB is also a file sharing protocol, not an IoT messaging one. Therefore, D, MQTT, is the correct answer.
Play from 2:11:48 Question 91. A mobile app team is ready to push a new build to production. Which release approach keeps the rollout safe while avoiding extra load on shared network capacity? Here are the options. Option A. Shadow. Option B. Rolling. Option C. A/B testing. Option D. Blue-green. Pause the video now if you want time to think about the answer.
The correct answer is B. Rolling. A rolling release swaps old instances for new ones in small batches, so extra capacity is never needed all at once and traffic stays steady. Blue-green runs a full second environment and shadow testing duplicates live traffic, so both add load. A/B testing is about comparing user behavior, not about a safe rollout. Now, why are the other options incorrect? Option A is incorrect. Shadow deployments mirror production traffic to the new version, which increases load. Option C is incorrect. A/B testing compares versions on users to measure results, rather than rolling out safely. Option D is incorrect. Blue-green needs a complete duplicate environment, consuming extra resources. Therefore, B, Rolling, is the correct answer.
Play from 2:13:21 Question 92. A cloud provider wants to prove to prospective customers that it handles their data securely, using an independent audit report. Which standard is most closely tied to that kind of audit? Here are the options. Option A. ISO 27001. Option B. ITIL. Option C. CIS. Option D. SOC 2. Pause the video now if you want time to think about the answer.
The correct answer is D. SOC 2. A SOC 2 report is produced by an independent auditor and describes how a service provider manages customer data against trust criteria such as security and privacy. Customers often request it when evaluating a provider. Now, why are the other options incorrect? Option A is incorrect. This is a certifiable management system standard, not an auditor's report on service controls. Option B is incorrect. This is a framework for IT service management, not a data protection audit. Option C is incorrect. These are hardening benchmarks for configuring systems, not a provider audit procedure. Therefore, D, SOC 2, is the correct answer.
Play from 2:14:52 Question 93. A gaming community hosts a worldwide chat board. Its monitoring shows the landing page suddenly responding slowly, yet internal checks show no faults or recent changes. What is the most probable cause? Here are the options. Option A. Phishing. Option B. DDoS. Option C. Human error. Option D. Cryptojacking. Pause the video now if you want time to think about the answer.
The correct answer is B. DDoS. Slow responses on a public page when nothing changed inside suggests the cause is outside: a flood of requests from many sources swamping the service. That is a distributed denial of service attack. The others do not produce sudden slowness for users worldwide. Now, why are the other options incorrect? Option A is incorrect. Phishing tricks people into giving up credentials and does not slow a site. Option C is incorrect. A staff mistake would appear as an internal change or misconfiguration, which monitoring would have caught. Option D is incorrect. Cryptojacking would show up as high CPU on the hosts, which internal monitoring would flag. Therefore, B, DDoS, is the correct answer.
Play from 2:16:27 Question 94. A site reliability engineer signs in to Grafana. What is the engineer most likely about to do? Here are the options. Option A. Monitor the health of a running application. Option B. Build a container image for an application. Option C. Release a new application version. Option D. Upload fresh code to a Git repository. Pause the video now if you want time to think about the answer.
The correct answer is A. Monitor the health of a running application. Grafana is a dashboard tool that charts metrics and logs, so people open it to see how systems are performing. It is not used to build containers, ship releases or store code. Now, why are the other options incorrect? Option B is incorrect. Containers are built with tools like Docker, not Grafana. Option C is incorrect. Releases go through deployment or CI/CD tools. Option D is incorrect. Code is pushed with Git clients, not Grafana. Therefore, A, Monitor the health of a running application, is the correct answer.
Play from 2:17:50 Question 95. Which cloud concept moves content closer to users in a given area so it loads faster? Here are the options. Option A. Availability zone. Option B. Edge computing. Option C. Region. Option D. Cloud bursting. Pause the video now if you want time to think about the answer.
The correct answer is B. Edge computing. To cut latency, content is placed physically near users, which is the job of edge locations. Regions and availability zones are broad hosting locations for workloads, not delivery points close to users, and bursting is about overflow capacity. Now, why are the other options incorrect? Option A is incorrect. An availability zone provides fault isolation within a region, not proximity for content. Option C is incorrect. A region is a large geographic area for hosting, too coarse to serve content near each user. Option D is incorrect. Cloud bursting shifts overflow demand to a public cloud and is not about location. Therefore, B, Edge computing, is the correct answer.
Play from 2:19:18 Question 96. A startup wants to run its own application code in the cloud without setting up any servers or other infrastructure. Which service model suits it best? Here are the options. Option A. XaaS. Option B. IaaS. Option C. SaaS. Option D. PaaS. Pause the video now if you want time to think about the answer.
The correct answer is D. PaaS. When a team only wants to push its code and let the provider handle servers, storage and networking, a platform service is the match. Developers focus entirely on the application. Software services are finished products, infrastructure services require setting up machines, and XaaS is just a broad umbrella term. Now, why are the other options incorrect? Option A is incorrect. XaaS is a general label for anything delivered as a service, not a specific model for hosting code. Option B is incorrect. IaaS requires provisioning and managing virtual machines and networks. Option C is incorrect. SaaS delivers a finished application and does not host your own code. Therefore, D, PaaS, is the correct answer.
Play from 2:20:54 Question 97. A DevOps engineer is setting up a new service that must reach the company's public cloud resources. What is the best way to authenticate that service? Here are the options. Option A. MFA. Option B. Cookie. Option C. API key. Option D. Username and password. Pause the video now if you want time to think about the answer.
The correct answer is C. API key. Applications need a non-interactive way to prove who they are. An API key can be issued per application, limited in scope, and rotated, so it suits this job. Methods that rely on a human being present are a poor fit. Now, why are the other options incorrect? Option A is incorrect. MFA requires a person to respond to a prompt, which a service cannot do. Option B is incorrect. Cookies keep browser sessions alive and are not a credential for an application. Option D is incorrect. Human-style credentials are harder to scope and rotate and are not the best practice for services. Therefore, C, API key, is the correct answer.
Play from 2:22:22 Question 98. A platform engineer tries to push changes straight to a protected branch and gets an error. What should the engineer do instead? Here are the options. Option A. Rebasing the branch. Option B. Merging the branch. Option C. Registering another remote. Option D. Opening a pull request for review. Pause the video now if you want time to think about the answer.
The correct answer is D. Opening a pull request for review. Protected branches are locked so nobody can change them directly. The expected route is to put the changes on a separate branch and open a pull request, which lets teammates review and then merge through the rules. The alternatives still try to alter or sidestep the protection. Now, why are the other options incorrect? Option A is incorrect. Rebasing rewrites history on that branch, which protection typically blocks too. Option B is incorrect. A direct merge is still a direct change to the protected branch, so it is also blocked. Option C is incorrect. A new remote only adds a repository location and does not get around the restriction. Therefore, D, Opening a pull request for review, is the correct answer.
Play from 2:23:56 Question 99. Which scenario best illustrates refactoring? Here are the options. Option A. Operating system upgrade. Option B. Setting up a registry for container images. Option C. Converting files into a safer format. Option D. Modifying application code so a platform can host it. Pause the video now if you want time to think about the answer.
The correct answer is D. Modifying application code so a platform can host it. Refactoring means changing an application's code so it works well on a new target, here a hosting platform. The other choices describe infrastructure, operating system, or file-level tasks that leave the application code alone. Now, why are the other options incorrect? Option A is incorrect. Upgrading an OS is an infrastructure change, not a code change. Option B is incorrect. A registry is a storage service for images and involves no change to the application's code. Option C is incorrect. Changing file formats concerns data, not application code. Therefore, D, Modifying application code so a platform can host it, is the correct answer.
Play from 2:25:23 Question 100. A network engineer needs a solution that is highly flexible, centrally managed, controllable through code, and able to adapt automatically as traffic changes. Which approach fits best? Here are the options. Option A. Network load balancer. Option B. Virtual private network. Option C. Content delivery network. Option D. Software-defined network. Pause the video now if you want time to think about the answer.
The correct answer is D. Software-defined network. Centralized control, programmable behavior, and automatic adjustment to traffic are the hallmarks of software-defined networking, where a controller manages the network through software. Load balancers, VPNs and content delivery networks each handle a narrower task. Now, why are the other options incorrect? Option A is incorrect. A load balancer spreads requests across servers but is not a centrally programmable network fabric. Option B is incorrect. A VPN builds encrypted tunnels and does not adapt to traffic. Option C is incorrect. A CDN caches content near users and does not provide central network control. Therefore, D, Software-defined network, is the correct answer.
