Role interview guide
DevOps engineer interview questions and what to expect
DevOps and site reliability interviews usually cover Linux, networking, CI/CD, containers and at least one cloud platform, and lean heavily on troubleshooting: you're described a broken or slow system and asked how you'd find the problem. Candidates commonly report scripting tasks, questions on infrastructure as code and monitoring, and for SRE roles, discussion of incidents, on-call and reliability targets.
What candidates commonly report; it varies by employer, level and country. Last reviewed 18 October 2026.
The usual stages
- 1Recruiter or HR screen
- 2Technical screen on Linux, networking and scripting, often with a short coding or shell task
- 3Troubleshooting round: talk through how you'd debug a failing service or server
- 4CI/CD, containers, Kubernetes and cloud discussion, often based on setups you've built
- 5Design round for experienced roles: a reliable deployment or infrastructure setup
- 6Hiring manager or behavioural round on incidents and teamwork
What interviewers check
- A methodical troubleshooting approach, from symptom to root cause
- Strong Linux and networking basics: processes, files, DNS, TCP and HTTP
- Automating instead of repeating manual steps: scripts, pipelines, infrastructure as code
- Monitoring and alerting that catches problems before users do
- Safe changes: rollbacks, gradual rollouts and least-privilege access
- Calm, blameless handling of incidents
Behavioral questions
Tell me about an outage you helped resolve. Walk me through the timeline.
A good answer covers
- How it was detected and how badly users were affected
- What you checked, in what order, and what turned out to matter
- The fix, the review afterwards, and what changed as a result
Describe a manual process you automated. What did it save?
A good answer covers
- The process and why it was slow or risky
- The script, pipeline or tool you built, and how you got people to use it
- Time saved or errors avoided, with a number
Tell me about a change you made that broke something. How did you handle it?
A good answer covers
- What you changed and what broke
- How quickly you noticed, rolled back and told people
- The safeguard you added: a review step, a test, a staged rollout
Tell me about a time developers pushed back on a process or tool you introduced.
A good answer covers
- Their concern, and whether it was fair
- How you adjusted the change or explained it better
- Whether it was adopted, and what you learned
Questions about the job
A website starts returning 502 errors. Talk me through how you'd find the cause.
A good answer covers
- Checking what changed recently and what the dashboards show
- Following the request path: load balancer or proxy, web server, application processes, health checks
- Reading logs, confirming the backend is up and reachable, then fixing or rolling back
A Linux server is running slowly. Which commands and checks would you start with?
A good answer covers
- CPU, memory, disk and network with tools such as top, free, df and iostat
- Finding the process responsible and reading its logs
- Checking for a full disk, heavy swapping or too many open connections
Explain how DNS resolution works, and how you'd debug a 'site can't be reached' report.
A good answer covers
- Resolver, root, top-level domain and authoritative servers, and caching with TTLs
- Using dig or nslookup to see what's actually returned
- Common causes: a wrong record, an expired domain, old answers cached after a change
Walk me through a CI/CD pipeline you've built or used, from a commit to production.
A good answer covers
- Build, test, security scan and packaging stages
- How it deploys: environments, approvals and rollout strategy
- How a bad release is caught and rolled back
What's the difference between a container and a virtual machine, and what does Kubernetes add on top of containers?
A good answer covers
- Containers share the host's kernel; virtual machines run their own operating system
- What Kubernetes adds: scheduling, restarts, scaling and service discovery
- When it's more than a small team needs
A pod in Kubernetes keeps restarting. How would you find out why?
A good answer covers
- kubectl describe and logs, including the previous container's logs
- Common causes: failing health checks, running out of memory, bad configuration or missing secrets
- Fixing the cause rather than only raising limits
Why use infrastructure as code, and how do you make changes to it safely?
A good answer covers
- Repeatable, reviewable infrastructure with a tool such as Terraform
- Reviewing the plan before applying, and handling state carefully
- Separate environments, and avoiding manual changes that drift from the code
Compare blue-green, canary and rolling deployments. When would you use each?
A good answer covers
- How each moves traffic to the new version
- How fast you can roll back, and the extra cost of each
- The metrics that decide whether a canary carries on
What would you monitor for a web service, and how do you avoid alerts that nobody acts on?
A good answer covers
- Latency, traffic, errors and saturation, plus a business signal such as orders
- Alerting on what users feel, with a clear runbook for each alert
- Reviewing noisy alerts regularly and tuning or removing them
Explain SLIs, SLOs and error budgets, and how they change what a team works on.
A good answer covers
- A measured indicator, a target for it, and the shortfall the target allows
- Using the budget to balance new features against reliability work
- An example target for a service you know
How would you set up a web app's infrastructure so it survives a server, or a whole zone, going down?
A good answer covers
- Several instances behind a load balancer, spread across zones
- A database with replicas, backups and restores you've actually tested
- Health checks, auto-scaling, and what would still fail if a whole region went down
How do you manage secrets and access permissions in the cloud?
A good answer covers
- A secrets manager or vault, never code or plain files committed to Git
- Least-privilege roles for people and services, and rotating keys
- Audit logs of who accessed what
Motivation and fit
Why DevOps or SRE? What do you enjoy about this kind of work?
A good answer covers
- What draws you: automation, reliability, helping developers ship faster
- An example of work you enjoyed
- An honest view of on-call and how you handle it
How would you spend your first month getting to know our systems?
A good answer covers
- Reading architecture notes, runbooks and recent incident reviews
- Shadowing on-call and starting with small, safe changes
- Spotting one repetitive task to automate and agreeing it with the team
Tips
- Practise troubleshooting answers in a fixed order: what changed, what the metrics show, then layer by layer.
- Know the commands you'd actually type; interviewers often ask for them by name.
- Build a small setup of your own: a container, a pipeline and a cloud deployment you can describe end to end.
- Whenever you describe a change, mention its blast radius and how you'd roll it back.
- If you hold a cloud certification, expect questions that check you've used the services, not just studied them.
- Describe incidents without blaming people; focus on the system and what changed.
Practise it out loud
A mock interview asks what devops engineer interviews usually test, follows up on your answers and gives you a report. The warm-up is free, needs no account and checks five of these questions as you answer.
About this guide
Written by Orbit from what candidates and interviewers commonly describe for this kind of job. The questions are Orbit's own practice questions, not questions taken from any employer's interviews. Processes change, so check the job post and what your recruiter tells you.