The scenario round
Expect something like: "The site is down. You are on call. Walk me through what you do." This is the round that matters, and it rewards a method rather than a guess.
- Establish impact first. What is broken, for whom, and how badly. Everything else follows from scope.
- Communicate. Say that you would post in the incident channel and set expectations. Candidates who debug silently lose points - in a real incident that is the actual failure.
- Mitigate before diagnosing. Roll back, fail over, scale up. Restoring service beats understanding the cause, and saying so explicitly is a strong signal.
- Then narrow it down. What changed recently? Deploys, config, infrastructure, certificates, a dependency.
- Check the obvious boring things - disk, memory, certificate expiry, DNS, a full connection pool.
- Afterwards: a blameless postmortem and one concrete action item.
CI/CD questions
- Design a pipeline for a service with multiple environments. What gates exist between them?
- Blue-green versus canary versus rolling - and when each is appropriate.
- How do you handle database migrations in a zero-downtime deploy? Expect follow-ups; this is a genuinely hard problem and interviewers know it.
- How do you manage secrets in a pipeline, and what is wrong with environment variables in the repo?
- A deploy succeeded but the service is broken. What should the pipeline have caught?
Kubernetes and containers
- A pod is in CrashLoopBackOff. Walk through diagnosing it.
- Requests versus limits, and what happens when you set limits too low - or omit them.
- Liveness versus readiness probes, and the failure mode of confusing them.
- How does a Service route traffic to pods, and what happens during a rolling update?
- What is in your container image that should not be, and why does image size matter?
Infrastructure as code and reliability
- Terraform state - where does it live, and what happens when two engineers apply at once?
- How do you handle drift between declared and actual infrastructure?
- What is the difference between monitoring and observability, in practical terms?
- What makes a good alert? The expected answer is actionable and rare; alerts nobody acts on are worse than none.
- Define an SLO for a service, and explain what the error budget is for.
How depth scales
Junior loops stay on tools and definitions. Mid-level loops live in the scenario round. Senior loops ask about organisational things - how you got teams to adopt a platform, how you reduced change failure rate, what you chose not to automate and why.
Where CoPilot Interview fits
Disclosure: we build an interview assistant, so treat this as a disclosed pitch. CoPilot Interview is a native Windows and macOS app that puts suggested answers on your own screen in about two seconds, across coding, system design and behavioral rounds, with a permanent free tier that needs no credit card.
There is a role page at DevOps interview help.
Frequently asked questions
What questions are asked in a DevOps interview?
Mostly scenario-based: an incident walkthrough, CI/CD pipeline design, Kubernetes debugging such as CrashLoopBackOff, infrastructure-as-code questions around Terraform state and drift, and reliability topics like SLOs and what makes a good alert. Tool trivia appears but rarely decides the outcome.
How do you answer a 'the site is down' interview question?
Establish impact and scope first, say you would communicate in the incident channel, then mitigate before diagnosing - roll back or fail over, because restoring service beats understanding the cause. Then look at what changed recently, starting with deploy history, and check boring causes like disk, certificates and DNS.
What Kubernetes questions come up in DevOps interviews?
Diagnosing CrashLoopBackOff, the difference between requests and limits and what happens when limits are wrong, liveness versus readiness probes, how Services route traffic during a rolling update, and what should not be in a container image.
How are senior DevOps interviews different?
Junior loops test tools and definitions, mid-level loops centre on the incident scenario, and senior loops move to organisational questions - driving platform adoption, reducing change failure rate, and what you deliberately chose not to automate.
Walk into the round with a second pair of eyes
CoPilot Interview is a native Windows and macOS app that puts suggested answers on your own screen in about two seconds, across coding, system design and behavioral rounds. The free tier is permanent: 15 minutes a day, no credit card.
Try it free