Site Reliability Engineer interview questions and answers.

Questions

Technical questions to practise out loud

1. 01

   How do SLOs and error budgets work?

   An SLO sets a target for a user-centered service indicator over a window; the error budget is the allowed unreliability. Burn rate guides alerts and whether teams can increase release risk or must restore reliability.

   Read the full breakdown
2. 02

   How do you triage a service incident as an SRE?

   Quantify affected users and SLO impact, assign incident command, inspect golden signals and recent changes, form and test one hypothesis at a time, mitigate safely before root cause, and communicate on a cadence.

   Read the full breakdown
3. 03

   How do you approach capacity planning?

   Forecast demand from historical growth and planned events, measure resource per unit of load, model peak and failure headroom, identify the first saturating dependency, load-test assumptions, and set scaling lead times.

   Read the full breakdown
4. 04

   How do metrics, logs, and traces work together?

   Metrics reveal trends and alert conditions, traces connect latency across a request, and logs explain discrete events. Correlation IDs and consistent service metadata let engineers move among them without duplicating every signal.

   Read the full breakdown
5. 05

   How do you make an on-call rotation sustainable?

   A sustainable rotation has actionable symptom-based alerts, manageable frequency, documented playbooks, safe access and automation, clear escalation, protected recovery time, regular training, and engineering work to remove recurring toil.

   Read the full breakdown
6. 06

   What makes a useful incident postmortem?

   A useful blameless postmortem gives impact and timeline, contributing technical and organizational conditions, detection and response analysis, evidence-backed causes, and a small set of owned actions that reduce recurrence or blast radius.

   Read the full breakdown
7. 07

   How do you test service reliability?

   Test reliability with load and stress tests, dependency-failure injection, zone or instance loss, timeout and retry validation, backup restores, and game days; define steady-state signals and safe abort conditions first.

   Read the full breakdown
8. 08

   How do you reduce alert fatigue?

   Remove alerts that are not urgent or actionable, alert on user symptoms and SLO burn, deduplicate correlated failures, tune windows from incident evidence, route ownership correctly, and review noisy pages as defects.

   Read the full breakdown
9. 09

   How do you plan disaster recovery?

   Define service-specific RTO and RPO, map dependencies and data, choose backup or multi-region recovery accordingly, automate rebuild and failover, protect credentials, and regularly restore and operate from the recovery environment.

   Read the full breakdown
10. 10

    How do you assess release reliability?

    Assess release reliability with change failure rate, rollback rate, canary error and latency deltas, SLO burn, incident linkage, and time to detect and recover; compare equivalent traffic and keep release markers in telemetry.

    Read the full breakdown