A.Deploy one service without redeploying the whole system
B.Deploy without writing any code
C.Deploy all services as a single unit
D.Deploy only once per year
Correct Answer: Deploy one service without redeploying the whole system
Explanation:
Independent deployment lets teams update and release individual services without affecting the entire application.
Incorrect! Try again.
19Which of the following is a major cloud provider offering DevOps services?
Intro to DevOps on AWS/GCP/Azure (conceptual only)
Easy
A.Amazon Web Services (AWS)
B.Adobe Illustrator
C.VLC Media Player
D.Microsoft Word
Correct Answer: Amazon Web Services (AWS)
Explanation:
AWS, along with GCP and Azure, is a major cloud provider offering a range of DevOps tools and services.
Incorrect! Try again.
20What does the abbreviation GCP stand for?
Intro to DevOps on AWS/GCP/Azure (conceptual only)
Easy
A.General Compute Program
B.Graphic Cloud Processor
C.Google Cloud Platform
D.Global Container Provider
Correct Answer: Google Cloud Platform
Explanation:
GCP stands for Google Cloud Platform, one of the major public cloud providers supporting DevOps workflows.
Incorrect! Try again.
21A company runs 200 containers across 15 hosts. When one host fails, the containers on it must be automatically restarted elsewhere without manual intervention. Which capability of a container orchestrator directly addresses this requirement?
Need for Orchestration
Medium
A.Container image signing
B.Self-healing and rescheduling
C.Build pipeline triggering
D.Image layer caching
Correct Answer: Self-healing and rescheduling
Explanation:
Orchestrators continuously monitor container/host health and reschedule failed workloads onto healthy nodes, providing self-healing. Image caching, signing, and pipeline triggers are unrelated to recovering from host failure.
Incorrect! Try again.
22As container counts grow, teams struggle to manually balance load, place containers, and scale services. Which set of problems is orchestration primarily designed to solve?
Need for Orchestration
Medium
A.Designing relational database schemas
B.Source code versioning and merging
C.Writing Dockerfiles and base images
D.Scheduling, scaling, and service discovery
Correct Answer: Scheduling, scaling, and service discovery
Explanation:
Orchestration platforms automate placement (scheduling), scaling up/down, and locating services (service discovery). Version control, image authoring, and schema design are outside orchestration's scope.
Incorrect! Try again.
23In a Kubernetes cluster, which control plane component is responsible for assigning newly created Pods to suitable nodes based on resource requirements and constraints?
Kubernetes Architecture Overview
Medium
A.etcd
B.kubelet
C.kube-proxy
D.kube-scheduler
Correct Answer: kube-scheduler
Explanation:
The kube-scheduler watches for unscheduled Pods and selects an appropriate node. The kubelet runs Pods on a node, kube-proxy handles networking rules, and etcd stores cluster state.
Incorrect! Try again.
24A developer changes cluster state via kubectl apply. Which component receives this request first and serves as the single entry point to the control plane?
Kubernetes Architecture Overview
Medium
A.kubelet
B.container runtime
C.kube-apiserver
D.kube-controller-manager
Correct Answer: kube-apiserver
Explanation:
The kube-apiserver is the front end of the control plane; all clients and components communicate through it. Controllers, kubelet, and the runtime act on state after it is stored via the API server.
Incorrect! Try again.
25Where does a Kubernetes cluster persistently store all of its configuration and state data as key-value pairs?
Kubernetes Architecture Overview
Medium
A.etcd
B.kube-proxy
C.the container registry
D.kubelet cache
Correct Answer: etcd
Explanation:
etcd is the consistent, distributed key-value store that holds all cluster state. kube-proxy manages network rules, the kubelet manages node workloads, and the registry stores images.
Incorrect! Try again.
26Two containers in the same application must share the same network namespace and communicate over localhost. What is the correct way to deploy them in Kubernetes?
Basic Kubernetes Concepts: Pods
Medium
A.Run them as separate DaemonSets
B.Use two Deployments linked by a Service
C.Create two separate Pods on the same node
D.Place both containers in the same Pod
Correct Answer: Place both containers in the same Pod
Explanation:
Containers within a single Pod share the same network namespace and can reach each other via localhost. Separate Pods or Deployments each get their own network identity, so localhost would not work.
Incorrect! Try again.
27Why is a Pod, rather than an individual container, considered the smallest deployable unit in Kubernetes?
Basic Kubernetes Concepts: Pods
Medium
A.A Pod is a snapshot of a container image
B.A Pod can only ever hold exactly one container
C.A Pod groups one or more tightly coupled containers sharing storage and network
D.A Pod represents a physical machine in the cluster
Correct Answer: A Pod groups one or more tightly coupled containers sharing storage and network
Explanation:
Kubernetes schedules Pods, which can contain one or more containers that share networking and storage. Pods are not limited to one container, are not physical machines, and are not image snapshots.
Incorrect! Try again.
28A team wants to update an application from version 1 to version 2 gradually, replacing old Pods with new ones while keeping the app available. Which Kubernetes object and strategy fits best?
Basic Kubernetes Concepts: Deployments
Medium
A.A single Pod recreated manually
B.A ConfigMap with a new value
C.A Service with session affinity
D.A Deployment with a rolling update
Correct Answer: A Deployment with a rolling update
Explanation:
Deployments support rolling updates that incrementally replace old Pods with new ones, avoiding downtime. Services route traffic, ConfigMaps store configuration, and a lone Pod offers no managed update strategy.
Incorrect! Try again.
29After a bad release, an operator wants to quickly return the application to its previous stable version. Which Deployment feature enables this?
Basic Kubernetes Concepts: Deployments
Medium
A.Manually deleting etcd entries
B.Rollback to a previous revision
C.Restarting kube-proxy
D.Editing the container image registry
Correct Answer: Rollback to a previous revision
Explanation:
Deployments keep a revision history, allowing kubectl rollout undo to revert to a prior working state. Editing etcd directly, restarting kube-proxy, or changing the registry does not restore application versions safely.
Incorrect! Try again.
30A Deployment specifies replicas: 4. Two Pods are accidentally deleted. What does the Deployment's controller do?
Basic Kubernetes Concepts: Deployments
Medium
A.Creates 2 new Pods to restore the desired count
B.Waits for manual recreation of the Pods
C.Deletes the remaining 2 Pods
D.Reduces the replica count to 2 permanently
Correct Answer: Creates 2 new Pods to restore the desired count
Explanation:
Deployments (via ReplicaSets) continuously reconcile actual state to the declared desired state of 4 replicas, recreating missing Pods automatically.
Incorrect! Try again.
31Pods are frequently created and destroyed, so their IP addresses change. Which Kubernetes object provides a stable endpoint to reach a group of Pods?
Basic Kubernetes Concepts: Services
Medium
A.A Service
B.A container image tag
C.A namespace
D.A Pod label
Correct Answer: A Service
Explanation:
A Service gives a stable virtual IP and DNS name that load-balances across matching Pods, decoupling clients from changing Pod IPs. Labels, image tags, and namespaces do not provide a stable network endpoint.
Incorrect! Try again.
32An application must be reachable from outside the cluster through a cloud provider's load balancer. Which Service type is most appropriate?
Basic Kubernetes Concepts: Services
Medium
A.LoadBalancer
B.ExternalName
C.ClusterIP
D.None (headless)
Correct Answer: LoadBalancer
Explanation:
The LoadBalancer Service type provisions an external load balancer via the cloud provider. ClusterIP is internal-only, headless disables load balancing, and ExternalName maps to an external DNS name.
Incorrect! Try again.
33How does a Service determine which Pods should receive its traffic?
Basic Kubernetes Concepts: Services
Medium
A.By matching Pod labels using a label selector
B.By matching container image names
C.By matching Pod IP addresses statically
D.By the alphabetical order of Pod names
Correct Answer: By matching Pod labels using a label selector
Explanation:
Services use label selectors to dynamically identify the set of Pods they route to. This lets Pods come and go while the selector keeps the Service endpoints current.
Incorrect! Try again.
34A large team finds that a single shared codebase forces all developers to coordinate every release, slowing delivery. Which microservices benefit most directly addresses this pain point?
Why Microservices?
Medium
A.Removal of the need for testing
B.Elimination of all network communication
C.Independent development and deployment of services
D.Guaranteed lower total infrastructure cost
Correct Answer: Independent development and deployment of services
Explanation:
Microservices let teams build, deploy, and release services independently, reducing coordination overhead. They typically add network communication and do not remove testing or guarantee lower cost.
Incorrect! Try again.
35One microservice needs a NoSQL database while another performs best with a relational database. How does a microservices architecture accommodate this?
Why Microservices?
Medium
A.Services cannot use databases directly
B.Only one database technology is allowed per cluster
C.Each service can choose its own technology stack
D.All services must share one database
Correct Answer: Each service can choose its own technology stack
Explanation:
Microservices support polyglot persistence and technology heterogeneity, letting each service pick the datastore and language best suited to its needs, unlike a monolith with a shared stack.
Incorrect! Try again.
36In a monolithic application, a small change to one module requires rebuilding and redeploying the entire application. What is the primary architectural reason for this?
D.All modules are tightly coupled in a single deployable unit
Correct Answer: All modules are tightly coupled in a single deployable unit
Explanation:
A monolith packages all functionality into one deployable artifact, so any change forces a full rebuild and redeploy. The other options describe microservice-style separation.
Incorrect! Try again.
37A payments component experiences 10x more load than the rest of an application. In a microservices architecture, what scaling advantage applies?
Monolithic and Microservice Architecture
Medium
A.Scaling requires rewriting the whole codebase
B.Only the payments service can be scaled independently
C.The entire application must be scaled together
D.Services cannot be scaled without downtime
Correct Answer: Only the payments service can be scaled independently
Explanation:
Microservices allow granular, independent scaling of just the high-demand service, using resources efficiently. A monolith would require scaling the whole application together.
Incorrect! Try again.
38A team deploys an update to the recommendation service without touching the checkout service, and users of checkout are unaffected. Which microservices benefit does this demonstrate?
Benefits (scalability, independent deployment)
Medium
A.Shared release cycles across all services
B.Independent deployment with fault isolation
C.Mandatory full-system redeployment
D.Elimination of inter-service dependencies
Correct Answer: Independent deployment with fault isolation
Explanation:
Deploying one service without impacting others illustrates independent deployment and isolation of change. It does not require shared releases or full redeployment, nor does it remove all dependencies.
Incorrect! Try again.
39If a service normally handling 500 requests/second must handle 2000 requests/second, and each instance handles 500 requests/second, how many instances are needed, and which property enables adding them?
Benefits (scalability, independent deployment)
Medium
A.8 instances, enabled by monolithic packaging
B.2 instances, enabled by vertical scalability
C.1 instance, enabled by caching
D.4 instances, enabled by horizontal scalability
Correct Answer: 4 instances, enabled by horizontal scalability
Explanation:
instances are required. Adding more instances is horizontal scaling, a key benefit of stateless microservices behind a load balancer.
Incorrect! Try again.
40A team wants a managed Kubernetes control plane so they do not have to operate master nodes themselves. Which pair of managed services fits this on AWS and Azure respectively?
Intro to DevOps on AWS/GCP/Azure (conceptual only)
Medium
A.Amazon S3 and Azure Blob Storage
B.Amazon EKS and Azure AKS
C.AWS Lambda and Azure Functions
D.AWS CloudTrail and Azure Monitor
Correct Answer: Amazon EKS and Azure AKS
Explanation:
EKS (AWS) and AKS (Azure) are managed Kubernetes services that operate the control plane for you. Lambda/Functions are serverless compute, S3/Blob are storage, and CloudTrail/Monitor are observability tools.
Incorrect! Try again.
41A team runs 200 containers across 15 hosts manually with shell scripts. During peak load, some containers crash and traffic is unevenly distributed. Which combination of orchestration capabilities is most essential to directly resolve both the crash-recovery and the traffic-distribution problems?
Need for Orchestration
Hard
A.Self-healing (automatic restart/reschedule) and load balancing
B.Static host assignment and manual port mapping
C.Log aggregation and centralized monitoring
D.Image caching and container layering
Correct Answer: Self-healing (automatic restart/reschedule) and load balancing
Explanation:
Crash recovery is solved by self-healing, which reschedules failed containers, while uneven traffic is solved by built-in load balancing. Logging and image caching help operations but do not directly restart containers or balance traffic.
Incorrect! Try again.
42In a Kubernetes control plane, if the etcd datastore becomes unreachable but existing worker nodes and their kubelets remain running, what is the most accurate immediate consequence?
Kubernetes Architecture Overview
Hard
A.The kubelets automatically promote themselves to control plane roles
B.Running Pods continue to serve traffic, but new scheduling and state changes fail
C.Services keep routing but Pods lose their persistent volumes instantly
D.All running Pods are immediately terminated across the cluster
Correct Answer: Running Pods continue to serve traffic, but new scheduling and state changes fail
Explanation:
etcd stores cluster state; losing it prevents new writes (scheduling, updates), but already-running Pods managed by kubelets keep running. Existing workloads are not immediately killed.
Incorrect! Try again.
43You define a Deployment with replicas: 3 and a Service of type ClusterIP selecting those Pods. One Pod is deleted manually via kubectl delete pod. What sequence correctly describes the system's response?
Basic Kubernetes Concepts: Pods, Deployments, Services
Hard
A.The ClusterIP changes to redirect traffic to a new node
B.The Deployment's controller creates a replacement Pod, and the Service updates its endpoints to include it
C.The Service recreates the Pod directly and notifies the Deployment
D.The Pod stays deleted because Services, not Deployments, own Pods
Correct Answer: The Deployment's controller creates a replacement Pod, and the Service updates its endpoints to include it
Explanation:
The Deployment (via its ReplicaSet) reconciles the desired replica count and spawns a new Pod. The Service's endpoint controller then adds the new Pod's IP once it matches the selector and becomes ready.
Incorrect! Try again.
44A monolithic e-commerce app is split into microservices. The Orders service must confirm payment via the Payments service before finalizing. Which challenge is introduced by this decomposition that did not exist in the monolith?
Monolithic and Microservice Architecture
Hard
A.Compile-time coupling between all modules
B.Shared in-process function call overhead
C.Distributed transaction consistency across service boundaries
D.Single-database schema locking on every write
Correct Answer: Distributed transaction consistency across service boundaries
Explanation:
In a monolith, a local ACID transaction spans both operations. Splitting into services introduces network boundaries, requiring patterns like Saga or eventual consistency to maintain correctness—a new distributed-transaction challenge.
Incorrect! Try again.
45A Pod contains two containers: an application container and a sidecar proxy. Which statement about their relationship inside the Pod is correct?
Basic Kubernetes Concepts: Pods, Deployments, Services
Hard
A.They must run on different nodes for isolation
B.They share the same network namespace and can communicate over localhost
C.Each container gets its own Pod IP and separate loopback interface
D.They cannot share mounted volumes
Correct Answer: They share the same network namespace and can communicate over localhost
Explanation:
Containers in the same Pod share the network namespace (one IP) and can reach each other via localhost and distinct ports. They can also share volumes, enabling the sidecar pattern.
Incorrect! Try again.
46A microservices system autoscales only the Search service during traffic spikes while other services stay at baseline. Which benefit does this specifically demonstrate that a monolith cannot easily achieve?
Benefits (scalability, independent deployment)
Hard
Microservices allow scaling individual components independently based on their own load. A monolith must scale the entire application as one unit, wasting resources on components that don't need scaling.
Incorrect! Try again.
47Which control plane component is responsible for deciding which node a newly created Pod should run on, based on resource requests and constraints?
Kubernetes Architecture Overview
Hard
A.kube-scheduler
B.kubelet
C.kube-proxy
D.kube-controller-manager
Correct Answer: kube-scheduler
Explanation:
The kube-scheduler watches for unscheduled Pods and binds them to suitable nodes using filtering and scoring (resources, affinity, taints). The kubelet then runs the Pod once assigned.
Incorrect! Try again.
48A Deployment update uses the default RollingUpdate strategy with maxUnavailable: 0 and maxSurge: 1 for 4 replicas. During the rollout, what is the maximum number of Pods (old + new) that can exist simultaneously?
Basic Kubernetes Concepts: Pods, Deployments, Services
Hard
A.3
B.5
C.8
D.4
Correct Answer: 5
Explanation:
With 4 desired replicas, maxSurge: 1 allows one extra Pod above desired (), and maxUnavailable: 0 ensures no old Pod is removed before a new one is ready, so the peak count is 5.
Incorrect! Try again.
49An organization adopts microservices primarily to allow different teams to use different tech stacks. Which term best captures this specific advantage?
Microservices communicate over language-agnostic APIs, letting each service choose the best-suited language or database—polyglot programming/persistence. This freedom is hard to achieve in a single monolithic codebase.
Incorrect! Try again.
50A team wants a managed Kubernetes control plane so they don't operate etcd or the API server themselves. Which mapping of managed Kubernetes services to cloud providers is entirely correct?
Intro to DevOps on AWS/GCP/Azure (conceptual only)
Hard
A.AWS → ECS, GCP → GKE, Azure → EKS
B.AWS → EKS, GCP → GKE, Azure → AKS
C.AWS → GKE, GCP → AKS, Azure → EKS
D.AWS → AKS, GCP → EKS, Azure → GKE
Correct Answer: AWS → EKS, GCP → GKE, Azure → AKS
Explanation:
The managed Kubernetes offerings are EKS (Elastic Kubernetes Service) on AWS, GKE (Google Kubernetes Engine) on GCP, and AKS (Azure Kubernetes Service) on Azure.
Incorrect! Try again.
51You need stable network access to a set of Pods from outside the cluster in a cloud environment, provisioning an external IP automatically. Which Service type is most appropriate?
Basic Kubernetes Concepts: Pods, Deployments, Services
Hard
A.Headless (clusterIP: None)
B.LoadBalancer
C.ExternalName
D.ClusterIP
Correct Answer: LoadBalancer
Explanation:
A LoadBalancer Service provisions a cloud provider's external load balancer with a public IP. ClusterIP is internal-only, ExternalName maps to a DNS name, and Headless returns Pod IPs directly without load balancing.
Incorrect! Try again.
52Consider declarative orchestration where you specify desired state rather than imperative steps. What is the core mechanism an orchestrator uses to continuously enforce this?
Need for Orchestration
Hard
A.A static configuration snapshot with no monitoring
B.A manual approval gate before every change
C.A reconciliation loop comparing actual state to desired state
D.A one-time provisioning script executed at deploy
Correct Answer: A reconciliation loop comparing actual state to desired state
Explanation:
Kubernetes controllers run continuous reconciliation loops: they observe actual state, compare it to the declared desired state, and take corrective actions to converge—the foundation of declarative orchestration.
Incorrect! Try again.
53For a small startup with 3 developers and an unproven product, which architectural choice is often recommended first, and why?
Monolithic and Microservice Architecture
Hard
A.A distributed monolith to combine both approaches
B.A serverless mesh of 50 functions for infinite scale
C.A monolith, to reduce operational complexity while the domain is still evolving
D.Microservices, to maximize independent deployment from day one
Correct Answer: A monolith, to reduce operational complexity while the domain is still evolving
Explanation:
The 'monolith-first' principle argues that premature microservices add distributed-system overhead. A monolith is simpler to build and refactor while service boundaries are still unclear, then decompose later.
Incorrect! Try again.
54On each worker node, which component maintains network rules to route Service traffic to the correct backend Pods?
Kubernetes Architecture Overview
Hard
A.kube-proxy
B.etcd
C.kube-apiserver
D.kube-scheduler
Correct Answer: kube-proxy
Explanation:
kube-proxy runs on every node and programs iptables/IPVS rules (or uses userspace mode) to forward traffic destined for a Service's virtual IP to one of its backend Pods.
Incorrect! Try again.
55Team A deploys a bugfix to the Notifications service without redeploying Orders or Billing. Which property enables this, and what is a key prerequisite to keep it safe?
Benefits (scalability, independent deployment)
Hard
A.Tight coupling; requires a shared database schema lock
B.Shared binary linking; requires recompiling all services
C.Independent deployability; requires backward-compatible API contracts
Correct Answer: Independent deployability; requires backward-compatible API contracts
Explanation:
Independent deployment lets one service ship without others. To avoid breaking consumers, changes must preserve backward-compatible interfaces (e.g., additive API changes, versioning).
Incorrect! Try again.
56A Pod repeatedly enters CrashLoopBackOff. Which explanation best describes what this state indicates about kubelet behavior?
Basic Kubernetes Concepts: Pods, Deployments, Services
Hard
A.The Service selector does not match the Pod labels
B.The image cannot be pulled from the registry
C.The container keeps failing on start, and kubelet restarts it with exponentially increasing delays
D.The Pod cannot be scheduled due to insufficient node resources
Correct Answer: The container keeps failing on start, and kubelet restarts it with exponentially increasing delays
Explanation:
CrashLoopBackOff means the container starts, crashes, and kubelet restarts it with a growing back-off delay. Scheduling failures show Pending, and image issues show ImagePullBackOff/ErrImagePull.
Incorrect! Try again.
57Conway's Law is often cited when justifying microservices. Which statement correctly applies it to service design?
Why Microservices?
Hard
A.System structure tends to mirror the communication structure of the organization that builds it
B.Microservices must always outnumber the teams building them
C.Code quality is inversely proportional to the number of services
D.Systems always fail at the network layer first
Correct Answer: System structure tends to mirror the communication structure of the organization that builds it
Explanation:
Conway's Law states that organizations design systems that copy their communication structures. Teams aligned to business capabilities naturally produce services around those capabilities, motivating team-scoped microservices.
Incorrect! Try again.
58A DevOps engineer wants a fully managed CI/CD pipeline service native to each cloud. Which grouping of native pipeline/build services is correct?
Intro to DevOps on AWS/GCP/Azure (conceptual only)
Hard
Native CI/CD tooling includes AWS CodePipeline/CodeBuild, Google Cloud Build, and Azure Pipelines (part of Azure DevOps). The other options list storage, compute, or database services, not CI/CD pipelines.
Incorrect! Try again.
59Which scenario represents the strongest justification for adopting a container orchestrator rather than plain Docker on a single host?
Need for Orchestration
Hard
A.Running hundreds of containers across many nodes needing automated scheduling, scaling, and failover
B.Building one Docker image for a static website
C.Running a single container locally for a developer demo
D.Storing configuration files in version control
Correct Answer: Running hundreds of containers across many nodes needing automated scheduling, scaling, and failover
Explanation:
Orchestration's value emerges at scale: multi-node scheduling, automated scaling, self-healing, and service discovery. A single local container or image build gains little from an orchestrator's overhead.
Incorrect! Try again.
60A system was split into microservices but every service shares one database and must be deployed together. This anti-pattern is best described as a:
Monolithic and Microservice Architecture
Hard
A.Well-designed microservice mesh
B.Pure event-driven system
C.Stateless serverless architecture
D.Distributed monolith
Correct Answer: Distributed monolith
Explanation:
A distributed monolith has services that are physically separated but tightly coupled (shared database, synchronized deploys), inheriting distributed-system complexity without the independence benefits of true microservices.
Incorrect! Try again.
Did this save you a night before the exam?
LPU Notes is free, and it stays free. Ads cover part of the server bill.
The rest comes out of a student's own pocket: the domain, the storage,
and keeping the site up through the weeks everyone needs it at once.
The payment button didn't load. An ad blocker or a filtered network is the usual reason.
to try again.
Nothing here is ever locked, and nothing unlocks. Chip in only if it was worth it.
What it pays for →