System Design

Observability & SLOs Study Portal

Metrics, logs and traces, RED and USE, cardinality cost, trace propagation, SLIs, SLOs and error budgets, burn-rate alerts that page on user pain rather than causes
Flash Cards Practice
System Design / Observability & SLOs / Key takeaways

Observability & SLOs · Key takeaways

1 min read
Senior14 min read
Rapid overview

Key takeaways

  • Metrics for dashboards and alerts, logs for exact detail, traces for where the time went; link them with trace ids.
  • RED for services, USE for resources; latency as percentiles from histograms.
  • Keep metric labels bounded; high-cardinality detail goes on traces and logs.
  • Propagate trace context over HTTP and messages; tail-sample to keep errors and slow requests.
  • SLI measures, SLO targets, SLA contracts; the error budget funds velocity.
  • Page on fast and sustained burn rate of user-facing SLOs, not on every cause.

See also