SRE Interview Preparation
ELI5 Explanation
SRE interviews test how you think during failures, not just what terms you memorize.
Technical Explanation
Expect layered questioning: fundamentals, system design tradeoffs, incident response behavior, and post-incident improvement. Your answers should tie business impact to technical decisions, and include measurement plans.
Visual
Hands-on Prep Commands
kubectl get pods -A
kubectl describe pod -n
kubectl logs -n --since=10m
kubectl get events -n --sort-by=.lastTimestamp
kubectl rollout undo deploy/ -n Debugging Scenario
Interviewer gives: "Latency tripled after a deploy, no obvious errors." Strong response: define impact, check SLO burn, inspect canary metrics, compare config/feature flags, identify rollback threshold, and communicate mitigation timeline.
Beginner
- What is SRE and how is it different from DevOps?
- Define SLI, SLO, and SLA with examples.
- What is an error budget?
- What are golden signals?
- What does blameless mean?
Intermediate
- How do you design an SLO for login service?
- How do burn-rate alerts reduce paging noise?
- How do you structure an on-call escalation policy?
- How do you run a useful postmortem?
- How would you plan DR for stateful workloads?
Scenario-based
- New release increases errors by 3%. Rollback or hotfix?
- Budget almost exhausted but product deadline is tomorrow. What do you recommend?
- Dependency outage lasts 4 hours. How do you protect user experience?
- Alert storm hides major outage. What immediate and long-term fixes?
- Leadership asks for five nines. How do you explain feasibility and cost?
Real-world Use Case
Candidates who quantify tradeoffs and show incident leadership usually perform better than candidates who only define terms. Interviewers look for ownership and calm reasoning under uncertainty.
Summary
You now have an end-to-end SRE foundation: reliability targets, observability, incidents, postmortems, DR, and chaos. Revisit scenarios regularly to sharpen production judgment.