DevOps Engineer
Interview Questions.
15 practice questions across 13 skills — each with what the interviewer is really listening for. Prepare with intent, not guesswork.
Free · No signup · Grouped by skill
How to use this page
Don't memorise answers. For each question, read what the interviewer is looking for, then practise answering out loud in your own words with a real example. The notes describe strong answers and common red flags — they are for your prep, not a script.
Linux
1 questionA production server is at 100% CPU. Which Linux commands do you reach for and in what order?
IntermediateWhat the interviewer is looking for: Strong answers move from top/htop to identify the process, then dig with ps, strace, or logs, forming hypotheses methodically. Randomly restarting services without diagnosis is a red flag.
Docker
1 questionExplain the difference between a Docker image and a container, and how layers affect image size.
BasicWhat the interviewer is looking for: Look for image as an immutable template and container as a running instance, plus using multi-stage builds and layer caching to keep images small. Confusing the two or bloated images with no cleanup are red flags.
Kubernetes
2 questionsWhat problem does Kubernetes solve that plain Docker does not?
IntermediateWhat the interviewer is looking for: Strong candidates cover orchestration, self-healing, scaling, service discovery, and rolling updates across many hosts. Treating Kubernetes as 'just Docker with extra steps' is a red flag.
Walk me through what happens when a Kubernetes pod keeps crashing (CrashLoopBackOff). How do you debug it?
AdvancedWhat the interviewer is looking for: Look for kubectl describe and logs, checking readiness/liveness probes, resource limits, and config/secrets. Only deleting and recreating the pod without reading logs is a red flag.
Jenkins
1 questionHow would you design a CI/CD pipeline in Jenkins for a service that must deploy multiple times a day?
IntermediateWhat the interviewer is looking for: Strong answers cover build, test gates, artifact versioning, and staged deploys with rollback, keeping pipelines as code. Manual deploy steps or no automated tests before deploy are red flags.
Terraform
1 questionWhat is Infrastructure as Code, and why is Terraform state so important?
IntermediateWhat the interviewer is looking for: Look for declarative, version-controlled infra and the role of state in mapping config to real resources, plus remote state and locking for teams. Editing infra manually alongside Terraform is a red flag.
AWS
1 questionHow do you keep cloud costs and security under control when provisioning AWS infrastructure?
AdvancedWhat the interviewer is looking for: Strong candidates mention least-privilege IAM, tagging, right-sizing, budgets/alerts, and avoiding public exposure by default. Wide-open security groups or long-lived root keys are serious red flags.
Git
1 questionHow do you structure Git branching and commits so infrastructure changes are safe to review?
BasicWhat the interviewer is looking for: Look for small reviewable changes, PRs with plan output, and protecting main. Pushing infra changes straight to main with no review is a red flag.
Prometheus
1 questionWhat signals do you put on a dashboard to know a service is healthy, and how do you avoid alert fatigue?
IntermediateWhat the interviewer is looking for: Strong answers reference the golden signals (latency, traffic, errors, saturation), alert on symptoms not causes, and tune thresholds. Alerting on everything until people ignore alerts is a red flag.
Incident Management
1 questionA deploy went out and error rates spiked. Walk me through your incident response.
AdvancedWhat the interviewer is looking for: Look for mitigating first (rollback/feature flag), communicating status, then a blameless post-mortem with action items. Debugging in prod while users suffer instead of rolling back is a red flag.
Bash
1 questionWhen would you write a Bash script versus a Python script for automation?
BasicWhat the interviewer is looking for: Strong candidates use Bash for simple glue around system commands and Python when logic, data structures, or error handling grow. Giant unmaintainable Bash scripts for complex logic are a red flag.
DevSecOps
1 questionHow do you handle secrets in a pipeline so they never end up in logs or version control?
AdvancedWhat the interviewer is looking for: Look for a secrets manager, injected env vars, masked logs, and rotation, never plaintext in repos. Hardcoding credentials or echoing them in CI output are serious red flags.
Load Balancing
1 questionExplain what happens at the network level when a request is slow: how do you tell if it is DNS, the load balancer, or the app?
AdvancedWhat the interviewer is looking for: Strong answers isolate each hop with timing, check health checks and target distribution, and use tracing. Guessing the layer without measuring is a red flag.
Collaboration
1 questionHow do you balance developer velocity with the guardrails DevOps is supposed to enforce?
IntermediateWhat the interviewer is looking for: Strong candidates treat DevOps as enabling teams with self-service and paved paths, not gatekeeping. Being the 'department of no' or removing all guardrails for speed are both red flags.
Behavioural & role-general
1 questionTell me about an outage or near-miss you were part of. What did you change afterward?
BasicWhat the interviewer is looking for: Look for honest reflection, a blameless mindset, and concrete systemic fixes (automation, guardrails, monitoring). Blaming an individual or claiming nothing ever breaks are red flags.