Opscraft
Interviews

Interview prep

Every topic has one interview question. Answer it out loud, then write it down and check it against what a strong answer covers. Sign in and AI feedback scores your written answer out of 10.

How to practise

  1. Say it out loud first, as if the interviewer just asked. Two minutes is about right.
  2. Structure it: what it is, how it works, a real example, the trade-offs, and what you'd watch for in production.
  3. Bring your own story. A guided project or a lab fix you did gives you something real to describe.
  4. Score yourself honestly against the strong-answer checklist, then try again a week later.

Foundations: Linux, scripting, Git & networking

  1. Explain to a friend with no technical background what happens when they open a shopping website, and point out where a DevOps engineer's work shows up along the way. DevOps from zero
  2. A Linux server is slow and users are complaining. Walk through how you'd troubleshoot it, with the commands you'd run. Linux
  3. Live exercise: write (talk through) a script that finds EC2 instances without an "owner" tag across all regions of an account and reports them, with an option to stop them. How do you make it production-safe? Python & Bash automation
  4. A teammate force-pushed to main and three merged PRs disappeared from the remote. Walk me through how you'd recover and what you'd change afterwards. Git
  5. Step by step, what happens when a user opens https://app.example.com hosted behind an AWS ALB in a VPC? Networking fundamentals
  6. Design subnets for VPC 10.0.0.0/16 with public, private and database tiers across 3 AZs. Explain your CIDR choices. IP addressing & subnetting
  7. Explain TCP vs UDP, and give two DevOps situations where the difference matters for troubleshooting or design. TCP vs UDP
  8. Explain the TCP handshake and teardown, and how you'd use tcpdump to tell a firewall drop from a closed port. The TCP handshake
  9. Explain the TLS handshake in your own words, and how you'd debug an expired or incomplete certificate chain in production. How SSL/TLS works
  10. Explain 502 vs 503 vs 504 behind Nginx or an ALB, and how you'd troubleshoot each. HTTP & status codes

DNS

  1. Explain everything a recursive resolver does to answer a query for api.example.com from a cold cache. How DNS works behind the scenes
  2. Explain A, CNAME, alias and MX records, and why you'd use an alias instead of a CNAME on AWS. DNS record types
  3. You need to move a production domain to a new load balancer with minimal impact. Explain your DNS TTL plan. DNS TTL & caching
  4. Explain the SOA record fields and how NS delegation from the registrar to Route 53 works. Name servers & SOA

Containers

  1. How would you make a Docker image smaller, faster to build and more secure for production? Docker
  2. Explain depends_on vs healthchecks in Compose and how you'd make an app wait for its database. Docker Compose
  3. Explain how Docker Swarm keeps a service running when a node dies, and what manager quorum means. Docker Swarm

AWS cloud

  1. Design the AWS foundation for a new customer-facing service on EKS with a PostgreSQL database: network, compute, data, secrets, HA/DR and cost controls. Walk me through your choices. AWS core services
  2. Your company has 30 AWS accounts, shared IAM users with access keys, and CI pipelines using stored keys. Lay out how you would move to a secure identity and access model. AWS IAM & multi-account security
  3. Explain public vs private hosted zones, and how an EC2 instance in a VPC resolves a private record. Route 53 hosted zones
  4. Design Route 53 routing for an app in two regions: active-active for latency, with failover if one region goes down. Route 53 routing policies
  5. Design a serverless pipeline that processes customer file uploads (up to 2 GB) into thumbnails and metadata, notifies users, and survives bad files and traffic spikes. Explain your failure handling, scaling limits and how you'd deploy and monitor it. Serverless with AWS Lambda

Infrastructure as code

  1. How do you structure Terraform for multiple environments, manage state and locking, and handle drift? Terraform
  2. Your company has 300 EC2 instances across 12 Auto Scaling groups and 4 EKS clusters, patched by hand. Design an image pipeline that keeps them patched and hardened, and explain how new images reach production safely. Image building with Packer

Web serving & caching

  1. Your company runs a web app behind Nginx on EC2 and the community ingress-nginx controller on EKS. Users report intermittent 502s and slow uploads failing. Walk through how you'd debug it, and how you'd plan the move off ingress-nginx. Nginx & reverse proxies
  2. An e-commerce site on AWS is slow during sales and the database struggles. Design the caching strategy from browser to database, and explain how you'd avoid serving stale prices or one user's data to another. Caching & CDNs

CI/CD & build tooling

  1. Design the CI/CD setup in GitHub Actions for 40 microservice repositories deploying to EKS across dev, staging and prod AWS accounts. How do you keep it consistent and secure? GitHub Actions
  2. Builds across your Java teams are slow, sometimes break with no code change, and nobody can say which library versions are in production. How would you fix the build and artifact setup? Build tools & artifact repositories
  3. How would you use a Makefile as the task runner for a containerised service, and how does CI use it? Makefile
  4. Describe the Jenkins controller/agent setup you'd run in production and how you'd scale and secure it. Jenkins
  5. What is a Jenkins shared library, how is it structured, and how would you roll out a breaking change to it? Groovy for pipelines
  6. Walk through the Jenkinsfile you'd write for build → test → deploy to staging → manual approval → prod. Jenkinsfile

Kubernetes, EKS, GitOps & releases

  1. Explain everything that happens from kubectl apply of a Deployment until the pod is serving traffic behind a Service. Kubernetes in depth
  2. Pods go into CrashLoopBackOff right after a release. Walk me through your troubleshooting and fix. Kubernetes troubleshooting
  3. Explain how pods get IPs on EKS, what causes IP exhaustion, and how IRSA lets a pod access S3 securely. Amazon EKS
  4. You run 30 microservices and 15 third-party components on EKS with Flux. Explain how you'd structure Helm and Kustomize in the GitOps repo, and how a change flows to production safely. Helm & Kustomize
  5. You're asked to roll out GitOps for 3 EKS clusters (dev, staging, prod) and 25 services. Describe the repo layout, how a commit reaches prod, how secrets work, and what you'd do when a bad change is merged. GitOps with Flux & Argo CD
  6. How would you introduce OPA Gatekeeper into an EKS cluster? Give two example policies and a safe rollout plan. Open Policy Agent
  7. Compare rolling, blue-green and canary deployments, and say which you'd use for a payments API and why. Deployment strategies
  8. Your platform runs 60 microservices on EKS owned by 12 teams. Security wants encryption and identity-based access between services, and teams want safer canaries. Would you adopt a service mesh, which one and how, and how would you avoid it becoming a source of outages? Service mesh in depth

Databases

  1. You're the DevOps engineer for a payments app on Aurora PostgreSQL. Describe your backup, high-availability and disaster-recovery design, how you'd handle an accidental data deletion, and how you'd run schema changes and a major version upgrade without downtime. Database operations

Observability & reliability

  1. You join a team with 3 EKS clusters, no consistent monitoring, and noisy CPU alerts that wake people up. Describe the observability platform you'd build and how you'd fix alerting. Observability: metrics, logs, traces
  2. Tell me about a production incident you handled end to end. What happened, what did you do, and what changed afterwards? SRE & incident management

Security

  1. Your CTO asks you to make the path from commit to production on EKS "secure by default" without slowing teams down. What do you put in place, in what order? DevSecOps & secrets management

Platform, cost & AI

  1. You're hired as the first platform engineer at a 150-engineer company with 25 teams, each running its own pipelines, Terraform and EKS manifests differently. Onboarding takes three weeks. How would you build an internal developer platform over the first year, and how would you know it's working? Platform engineering & Backstage
  2. Leadership says the AWS bill (mostly EKS, RDS, NAT and CloudWatch) grew 60% in six months while traffic grew 15%. You own the platform. Walk through how you'd investigate, what you'd change in the first 90 days, and how you'd keep costs under control afterwards. Cloud cost control (FinOps)
  3. Your company wants to launch an internal assistant that answers questions from company documents, and the data science team also needs to retrain a fraud model weekly. As the platform engineer, design the infrastructure on AWS and EKS for both, covering serving, GPUs, releases, monitoring, security and cost. Running AI/ML systems

Interview readiness

  1. Tell me about the most impactful piece of platform automation you've built, using the STAR format. Senior DevOps interviews

Optional extras

  1. Map an AWS architecture you know (VPC, EKS, IAM, S3, CloudFormation) to its Azure equivalents. Azure for AWS engineers