DevOps Engineer technical round: Linux, CI/CD and IaC

DevOps technical rounds sample broadly because the role is broad. Expect questions across these areas, often woven into a single scenario. Cloud Engineer / AWS Jobs at the same companies overlap heavily with this scope, especially on Linux debugging and infrastructure-as-code.

• Linux and networking: processes, permissions, systemd, DNS, and debugging a service that will not start.
• CI/CD: pipeline design, build stages, artefact promotion, and rollback strategy.
• Containers and orchestration: Docker fundamentals and Kubernetes concepts like pods, services and deployments.
• Infrastructure as code: Terraform or similar, state management, and idempotency.

The signal they want is a systematic troubleshooter, not someone who recites commands.

Scenario questions dominate. Practise narrating a diagnostic path.

'A deployment succeeded but the service returns 502. Walk me through debugging.'

They want a layered approach: is the pod running, are readiness probes passing, is the service routing to healthy endpoints, is the app listening on the expected port.

'The disk is full on a production host at 2am. What do you do?'

Show you stabilise first (find and clear the biggest offenders safely), then fix the root cause.

'A Terraform apply wants to destroy a database. Why, and what now?'

This probes state drift and safe practices.

Question: 'A pod keeps restarting. How do you diagnose it?'

Strong answer: 'I would check the pod status and events first, then the logs of the current and previous container. A CrashLoopBackOff usually means the process exits, so I would look at exit codes, then liveness probe settings, resource limits causing OOMKills, and config or secret issues. I would reproduce locally if possible before changing production.' It is layered and evidence-led.

Weak answer: 'I would delete the pod and let it recreate.' That masks the symptom without diagnosis and shows no method. Interviewers read it as someone who reacts rather than investigates, which is a liability on call.

Depth of method beats memorised commands.

• Keep Linux debugging sharp: logs, processes, ports, permissions, disk and memory.
• Be able to sketch a CI/CD pipeline with build, test, deploy and rollback stages.
• Practise explaining Kubernetes failure modes and Terraform state pitfalls out loud.

Because so much of this round is spoken troubleshooting, rehearse narrating your diagnostic path clearly. A free AI voice mock interview on InterviewPrep generates a DevOps mock from your CV and a real job description, then scores your reasoning delivery, pace and filler words, so your on-call thinking sounds calm and structured rather than scattered.