The short answer
Good DevOps interview questions ask how a candidate behaves when production breaks and when delivery slows, not which tools they know. Ask for the first ten minutes of an incident, how they would make a slow pipeline fast without losing safety, and how they roll back a bad deploy. Score clarifying questions, named failure modes and stated trade-offs.
Hiring for this?
Conntal places DevOps engineers on contract, contract-to-hire or direct hire, each screened by a practitioner before you see a CV.
Tool lists tell you what a DevOps engineer has touched. These questions tell you what they would do. They are built from the screen Ace8 and AceMQ practitioners use for Conntal's DevOps roles, generalised so any team can use them.
1. Walk me through the first ten minutes of an incident.
Our published DevOps screen uses a RabbitMQ quorum queue losing its leader during a Kubernetes node upgrade while the dashboard shows healthy. The senior answer separates the expected event (leader re-election) from the real bug (a dashboard that lies), and asks about the PodDisruptionBudget before restarting anything.
2. The pipeline takes forty minutes. Make it ten without losing safety.
Listen for measurement first — which stage is slow — then caching, parallelism and test selection. Weak answers delete tests.
3. A deploy is bad. How do you roll back, and what can't you roll back?
The strong answer separates code from data: a schema migration or a message already published cannot be rolled back by redeploying the old version.
4. How do you keep secrets out of the repository?
Look for a secrets manager, short-lived credentials and scanning in the pipeline — and an answer for what happens when a secret is committed anyway: rotate first, then clean history.
5. What would you put on the on-call dashboard for a new service?
Strong candidates start from user-facing symptoms — errors, latency, saturation — rather than every metric the platform exposes.
6. Infrastructure as code drifted from reality. What now?
Listen for finding out why it drifted before overwriting it, because the manual change may be the fix someone made during an incident.
7. Tell me about a change you made that broke production.
Everyone senior has one. Listen for ownership and what changed in the process afterwards.
Hiring for this?
Conntal places DevOps engineers on contract, contract-to-hire or direct hire, each screened by a practitioner before you see a CV.
Related questions
What questions should I ask in a DevOps engineer interview?
Ask scenario questions: the first ten minutes of an incident, speeding up a slow pipeline safely, what cannot be rolled back, handling a committed secret, and dealing with infrastructure-as-code drift.
How should DevOps interview answers be scored?
Score whether the candidate asked a clarifying question, named a failure mode and stated a trade-off. Tool knowledge alone does not separate senior from mid-level engineers.