Telemetry: By The Numbers
summary
Telemetry instruments are used to capture measurements and turn them into actionable metrics using Telemetry and telemetry_metrics, not just pageviews. The ecosystem provides five metric types: counter sum last_value summary distribution, and reporters (example: telemetry_metrics_statsd) perform aggregation and send data. Libraries like Plug, Ecto, Absinthe, Broadway emit events so apps only attach handlers or reporters. Good metrics are comparable, understandable, usually ratios (errors/requests, signups/churn), change behavior, and avoid vanity counts; example goals mentioned include 20,000 req/s and latency buckets <200/500/1000 ms.
AI-written summary · about
chapters (7)
- 0:00Introduction to telemetry
- 3:51Why monitoring matters
- 8:06Measurements vs metrics
- 13:50Avoiding vanity metrics
- 15:17Introducing telemetry library
- 21:21Setting up telemetry metrics
- 25:37Key takeaways and conclusion
related talks
39mTelemetry ...and metrics for all
42mOpenTelemetry Metrics 101
36mTelemetry and LiveDashboard
19mTelemetry is Not Having to Hope- 36mTelemetry and LiveDashboard
36mMetrics - What's your code actually doing?
30mPanel: Telemetry That Matters
02h 41mThe Science of Signals: Mastering Telemetry for Observability