Interview - DevOps Engineer

1

Easy

0 / 2 points

What is the difference between Docker and Kubernetes?

Docker: Container runtime platform for packaging and running applications in isolated containers

Kubernetes: Container orchestration platform for managing, scaling, and deploying containerized applications across clusters

Docker creates containers, Kubernetes manages multiple containers at scale

Kubernetes can use Docker as container runtime (among others like containerd, CRI-O)

2

Easy

0 / 2 points

Explain the basic Linux file permissions system (chmod, chown).

Read (r=4), Write (w=2), Execute (x=1) permissions for Owner, Group, Others

chmod: Changes file permissions (e.g., chmod 755 file.sh = rwxr-xr-x)

chown: Changes file ownership (e.g., chown user:group file.txt)

Examples: chmod +x script.sh (add execute), chmod 644 config.yaml (rw-r--r--)

3

Easy

0 / 2 points

What is the purpose of a Git branch and how do you create one?

Git branch: Isolated line of development for features/fixes without affecting main codebase

Create branch: git checkout -b feature-name or git branch feature-name

Switch branches: git checkout branch-name or git switch branch-name

Merge branch: git merge feature-name (merges into current branch)

4

Easy

0 / 2 points

What is Infrastructure as Code (IaC) and why is it important?

IaC: Managing infrastructure through code files instead of manual configuration

Benefits: Version control, reproducibility, consistency across environments

Tools: Terraform, Ansible, CloudFormation, Pulumi

Enables automation, reduces human error, faster provisioning, documentation as code

5

Easy

0 / 2 points

Explain the difference between REST and SOAP APIs.

REST: Lightweight, uses HTTP methods (GET/POST/PUT/DELETE), JSON/XML, stateless

SOAP: Protocol-based, XML only, built-in security (WS-Security), stateful operations possible

REST: Simpler, faster, better for web services, mobile apps

SOAP: More standardized, better for enterprise integrations, financial transactions

6

Medium

0 / 3 points

How do you write an Ansible playbook to install and configure Nginx on multiple servers?

Define inventory file with target hosts groups

Create playbook with tasks: install nginx package (apt/yum module), start/enable service

Use templates module for nginx.conf configuration with Jinja2 variables

Add handlers to restart nginx on config changes, use become: yes for sudo privileges

7

Medium

0 / 3 points

Explain the difference between KVM and VMware virtualization. When would you use each?

KVM: Open-source, Linux kernel-based, type-1 hypervisor, free, built into Linux

VMware: Commercial, mature ecosystem (vSphere, vCenter), advanced features, enterprise support

KVM: Better for Linux-heavy environments, cost-sensitive projects, custom solutions

VMware: Enterprise environments needing vendor support, Windows workloads, existing VMware infrastructure

8

Medium

0 / 3 points

How do you implement a CI/CD pipeline using Jenkins for a Python application?

Create Jenkinsfile: checkout code from Git, install dependencies (pip), run tests (pytest)

Build Docker image with application, tag with build number/commit hash

Push image to container registry (Docker Hub, ECR, Harbor)

Deploy to environment: update Kubernetes manifests, helm upgrade, or Ansible playbook trigger

9

Medium

0 / 3 points

What is a Kubernetes Pod vs Deployment vs Service? Explain each.

Pod: Smallest unit, one or more containers sharing network/storage, ephemeral

Deployment: Manages ReplicaSets, ensures desired number of Pods running, rolling updates, rollbacks

Service: Stable network endpoint (ClusterIP/NodePort/LoadBalancer) to access Pods

Use case: Deployment creates/manages Pods, Service provides stable access to those Pods

10

Medium

0 / 3 points

How do you troubleshoot a Linux server that has high CPU usage?

Use top/htop to identify processes consuming CPU, check %CPU column

Check system logs: journalctl, /var/log/syslog, dmesg for errors/warnings

Analyze process: ps aux, lsof -p PID (open files), strace (system calls)

Solutions: kill/restart rogue process, optimize application, add resources, check for malware/crypto miners

11

Medium

0 / 3 points

Explain Python virtual environments and how you use them in deployment.

Virtual environment: Isolated Python installation with its own packages, avoiding dependency conflicts

Create: python -m venv venv or virtualenv venv

Activate: source venv/bin/activate, install packages: pip install -r requirements.txt

Deployment: Use venv in Docker (COPY requirements.txt, RUN pip install), systemd services, or container images

12

Hard

0 / 4 points

Design a highly available Kubernetes cluster with disaster recovery. What components and strategies would you use?

Multi-master setup: 3+ master nodes with etcd cluster (quorum), external load balancer for API server

Worker nodes across multiple availability zones/regions for fault tolerance

Storage: Persistent volumes with replication (Ceph, Longhorn, cloud storage classes)

Backup: Velero for cluster backups, etcd snapshots, GitOps for config (ArgoCD, Flux), monitoring with Prometheus/Grafana

13

Hard

0 / 4 points

How would you implement zero-downtime deployment for a microservices application on Kubernetes?

Use rolling updates: strategy.type=RollingUpdate, maxSurge=1, maxUnavailable=0

Readiness probes: ensure new pods receive traffic only when ready

Blue-green deployment: two identical environments, switch traffic with Service selector

Canary deployment: Istio/Nginx Ingress for gradual traffic shifting (10%→50%→100%)

14

Hard

0 / 4 points

You have a legacy monolithic application. Explain your strategy to containerize and migrate it to Kubernetes.

Assessment: Analyze dependencies, databases, file storage, external integrations

Containerization: Create Dockerfile, extract configs to environment variables/ConfigMaps

Database migration: StatefulSet or external managed DB, use initContainers for schema

Phased rollout: Run parallel environments, migrate traffic gradually, implement health checks, monitoring

15

Critical

0 / 5 points

Design a complete monitoring and alerting system for a Kubernetes production environment with 50+ microservices. Include metrics, logs, traces, and SLA monitoring.

Metrics: Prometheus for scraping metrics, Grafana dashboards, ServiceMonitors for auto-discovery

Logging: EFK stack (Elasticsearch, Fluentd/Fluent-bit, Kibana) or Loki for log aggregation

Tracing: Jaeger or Tempo for distributed tracing, OpenTelemetry instrumentation

Alerting: AlertManager rules for critical thresholds, PagerDuty/Slack integration, SLO dashboards (error rate, latency)

Cost optimization: Metrics retention policies, log sampling, resource quotas

16

Critical

0 / 5 points

Implement a complete GitOps workflow for managing infrastructure and applications. Include branch strategy, approvals, and automated testing.

Git repository structure: separate repos/directories for infrastructure (Terraform) and apps (Helm charts, manifests)

Branch strategy: main (production), staging, feature branches with PR reviews

ArgoCD/Flux for continuous deployment: watches Git, syncs to clusters, health checks

Pre-merge validation: CI pipeline runs terraform plan, helm lint, kube-score for security/best practices

Approvals: CODEOWNERS for critical paths, automated tests must pass, manual approval for production

17

Critical

0 / 5 points

Design a disaster recovery plan for a multi-region Kubernetes deployment with RTO of 15 minutes and RPO of 5 minutes.

Multi-region active-active: Global load balancer (AWS Route53, Cloudflare), data replication between regions

Backup strategy: Velero with 5-minute schedule, etcd snapshots every 5 minutes, database replication (sync/async)

Automated failover: Health checks detect region failure, DNS switches traffic automatically

Runbooks: Automated recovery scripts, failback procedures, regular DR drills (chaos engineering)

Data consistency: Use distributed databases (CockroachDB, Cassandra) or conflict resolution strategies

18

Critical

0 / 5 points

You need to optimize costs for a Kubernetes cluster running 200+ microservices with varying load patterns. Describe your comprehensive approach.

Right-sizing: Use VPA (Vertical Pod Autoscaler) for recommendations, set proper requests/limits based on actual usage

Autoscaling: HPA (Horizontal Pod Autoscaler) for CPU/memory, KEDA for custom metrics, cluster autoscaler for nodes

Spot instances: Use spot/preemptible nodes with Pod Disruption Budgets for non-critical workloads (60-90% savings)

Resource optimization: Namespace quotas, PriorityClasses for critical pods, bin packing with node affinity

Cost monitoring: Kubecost for per-service cost attribution, alerts on budget thresholds, reserved capacity for base load

⚙️ DevOps Engineer

📋 Candidate Information

📋 Interview Report