From Noise To Signal: Building Smart Observability Pipelines With OpenTelemetry
summary
Live demo builds an OpenTelemetry pipeline that enriches Kubelet metrics with deployment and image labels, filters out readiness/health-check traces with OTTL, transforms HTTP targets to redact product IDs, and implements tail-based sampling by routing spans via trace ID from a daemonset to a sampling deployment. Back ends are Loki/Grafana/Tempo/Mimir and span‑metrics emit exemplars for metric→trace drilldown. The pipeline reduces noisy traces and storage cost in the demo and the changes apply without app code edits.
AI-written summary · about
chapters (12)
- 0:00Introduction
- 1:49OpenTelemetry fundamentals
- 3:32Cloud native observability challenges
- 7:25Collector architecture patterns
- 10:37Processor deep dive with OTTL
- 19:06Connectors and signal transformation
- 23:36Live demo setup
- 26:42Enriching signals with Kubernetes
- 30:55Filtering health check noise
- 34:26Masking PII in data
- 39:44Tail sampling for efficiency
- 45:12Q&A
related talks
28mEnd-to-end CI/CD observability with open source
17mFrom noise to signal: piping your way to clear observability
23mModern Application Debugging: An Intro to OpenTelemetry
50mGuide to Observability with OpenTelemetry
17mOpenTelemetry Looks Good To Me: Loki, Grafana, Tempo, Mimir
36mObservability For You and Me with OpenTelemetry
50mOpen Telemetry: The Rise of Open Standards in Observability
23mOpenTelemetry Auto Instrumentation for Effortless Observability