Behavioral (STAR method)
Why interviewers ask behavioral questions
They predict how you work under pressure, communicate, and learn — not whether you memorized kubectl flags.
Use STAR: Situation → Task → Action → Result (add Lesson if you improved a process).
High-yield prompts (DevOps / SRE / Platform)
| Theme | Example question |
|---|---|
| Incident | Tell me about a production outage you led or helped resolve. |
| Automation | Describe something repetitive you automated and the outcome. |
| Trade-off | When did you ship faster vs. make something more reliable? |
| Conflict | Disagreement with dev or security — how did you resolve it? |
| Failure | A migration or deploy that went wrong — what did you change after? |
| Mentoring | How do you help juniors with Kubernetes or CI/CD? |
Answer skeleton (2–3 minutes)
- Context — product, scale, team (30 sec).
- Your role — owner, IC, on-call (15 sec).
- Actions — specific steps, tools, comms (60–90 sec).
- Result — metrics: MTTR, error rate, cost, time saved (30 sec).
- Lesson — runbook, alert, guardrail, postmortem action item (15 sec).
Phrases that sound senior
- “We narrowed blast radius by …”
- “I drove the incident channel timeline and assigned workstreams.”
- “We added a pre-merge check so this class of bug can’t reach prod again.”
- “I wrote the runbook and trained the next on-call.”
Red flags to avoid
- Blaming only “the dev” or “the vendor” with no self-reflection.
- Vague results (“it got better”) with no numbers.
- Stories where you had no agency — pick another example.
Practice
Write 5 STAR stories on index cards: one outage, one automation win, one failure, one cross-team conflict, one prioritization call. Reuse them across companies with light tailoring.