Kubernetes Observability with OpenTelemetry
summary
Argues OpenTelemetry actually standardizes collection across logs, metrics and traces, with otel collector as DaemonSet and auto-instrumentation for six to seven languages to get data without changing code. Shows semantic conventions that tag Kubernetes resources so logs, metrics and spans share context (semantic conventions for k8s attributes). Warns about metric cardinality and cost with a real-world cardinality explosion ($300k example). Demos include Helm/operator installs, PromQL-based alerts and a visual collector config tool; beta platform is live.
AI-written summary · about
chapters (7)
- 0:00Introduction and background
- 4:53Observability pillars explained
- 10:12OpenTelemetry fundamentals
- 14:29Metrics and cardinality risks
- 21:12Semantic conventions and context
- 26:05Kubernetes observability setup
- 29:40Vendor lock-in and Q&A
related talks
50mOpen Telemetry: The Rise of Open Standards in Observability
47mWhat Is This OpenTelemetry Thing?
43mWhat is this OpenTelemetry thing?
50mGuide to Observability with OpenTelemetry
28mImplementing Observability in Kubernetes
46mObservability Ecosystem in Kubernetes
35mOpenTelemetry for Busy People
23mModern Application Debugging: An Intro to OpenTelemetry