Skip to content

Playlist - SRE, 1 hour

Seven reads in order, ~60 minutes. By the end you have working mental models for what SRE actually does day-to-day: define SLOs, detect incidents, contain blast radius, design for resilience, automate toil away.

The reads, in order

  1. Observability basics (~7 min) Logs, metrics, traces - the three pillars and what each is for. The foundation of SRE.

  2. Idempotency explained (~6 min) Retry-safe operations. The single design property that makes incident response (and resilient systems) possible.

  3. Eventual consistency (~7 min) Why distributed systems sometimes show stale data, and why "make it strongly consistent" is rarely the right answer in production.

  4. Architecture pattern: disaster recovery (~12 min) The four DR archetypes (backup/restore, pilot light, warm standby, multi-site active-active) with RTO/RPO mapping.

  5. Architecture pattern: chaos engineering (~10 min) The discipline of injecting failures to verify resilience holds. Hypothesis → measure steady state → inject → observe → fix.

  6. Architecture pattern: multi-region active-active (~8 min) The "always-on" shape. When to use it, when it's overkill.

  7. Architecture pattern: cell-based architecture (~8 min) Blast-radius reduction. The architectural answer to "why did one bad deploy take everything down?"

What you can do after this playlist

  • Define an SLO and explain what its error budget governs.
  • Pick a DR pattern given an RTO/RPO target and cost constraint.
  • Articulate why idempotent operations are the prerequisite for safe retries.
  • Identify single-point-of-failure shapes in an architecture diagram.
  • Discuss when chaos engineering is worth the operational risk and when it isn't.

Postmortem case studies for context

Read these alongside the playlist:

Next steps

If you want to build: - Set up a monitoring stack - Prometheus + Grafana + Alertmanager - Run a DR drill - tested failover, measured RTO/RPO

If you want to go deeper: - SRE and reliability topic index - everything in one place - Service comparison: observability + monitoring - Troubleshooting guides - per-cloud and Kubernetes

If you want a cert: - GCP Cloud DevOps Engineer (PDOE) - the most SRE-flavored cloud cert - AWS DevOps Engineer Pro (DOP-C02) - Azure DevOps Engineer Expert (AZ-400) - Kubernetes CKA - if you operate on K8s - DevOps / SRE roadmap

Required reading outside this repo