Observability; a gentle introduction
summary
Starts from a 3am on‑call example where an engineer gets paged with noisy alerts and an 80% CPU symptom that means nothing without context, then shows replacing dispersed metrics and logs with instrumented apps, traces and centralized pipelines using Prometheus and OpenTelemetry, measuring business SLIs like revenue per minute, and setting SLOs and SLIs so alerts only fire when impact exceeds tolerance. Claims observability is a system property not a tool, notes millions of scraped metrics can harm performance, and recommends Service Level Objects to align teams while enabling frequent deploys (10x/day) and faster fixes.
AI-written summary · about
chapters (8)
- 0:00Observability challenges
- 3:54Monitoring to observability history
- 9:15Monitoring problem and consequences
- 12:49Origins of observability
- 16:42Three pillars of observability
- 20:34SLIs SLOs and SLAs
- 23:57Instrumenting apps for metrics
- 26:30Observability tools and culture
related talks
13mObservability that matters
35mObservability: Building Confidence & Answering the What, Where, How
42mObservability can help you understand change
33mEffective observability in microservice architectures
40mDemystifying Observability and Monitoring
32mThe Role of Observability in Cloud Native Environments
06mIgnite: Observability Observed
39mThe 5W’s of Observability with Practical Examples