Observability for Developers: How to Get There
summary
Argues that developers gain more from observability than ops by shipping structured traces and metadata instead of just grepping logs. Shows Honeycomb dashboards with a custom event_time_delta_sec metric and ~1s query ingest SLA to confirm user complaints, uses feature flags captured per-request to compare slices (including a 10% rollout example), and describes capturing algorithm outputs to A/B which NP‑complete heuristic wins in production. Recommends gradual instrumentation, consistent naming, and tracing as the next step.
AI-written summary · about
chapters (11)
- 0:00Introduction and speaker intro
- 1:26Observability as a developer mindset
- 6:24From monitoring to curiosity-driven exploration
- 9:15Debugging with production data
- 12:56Understanding impact before shipping
- 15:08Testing in production with real data
- 18:19Feature flags and observability
- 20:26Bridging dev and ops with shared tools
- 23:18Instrumentation and structured data
- 27:32From logs to tracing and analysis
- 30:04Tracing as developer-friendly observability
related talks
29mObservability for Developers: How to Get From Here to There
44mObservability: Superpowers for Developers
50mObservability and the Development Process
33mWhat Developers Should Know About Observability
23mA peek into observability from a tester's lens
45mObservability is for User Happiness
35mObservability: Building Confidence & Answering the What, Where, How
29mObservability; a gentle introduction