OpenTelemetry Looks Good To Me: Loki, Grafana, Tempo, Mimir
summary
Two small Java apps run via docker-compose and send metrics, traces and logs through an collector-contrib gateway to Loki, Tempo and Mimir with Grafana dashboards. Tracing uses a parent-based 10% trace sampling default to capture full traces; head sampling is recommended for small setups, tail sampling for centralized control. Demo generates ~8–10% errors, root cause surfaces as relation b table does not exist, dashboards link from RED metrics to the error traces, fixes drop noisy traces to under 2%. Logs sent via OTLP/gRPC or file scraping.
AI-written summary · about
chapters (7)
- 0:00OpenTelemetry intro
- 1:47Demo setup overview
- 3:15RED method for monitoring
- 6:21Tracing strategies
- 9:32Optimizing sampling
- 11:57Live error/trace demo
- 14:20Conclusion
related talks
35mOpenTelemetry for Busy People
01h 02mThe Power of Metrics, Logs & Traces with Open Source
32mOpenTelemetry feat. logs, metrics and traces
45mFrom Noise To Signal: Building Smart Observability Pipelines With OpenTelemetry
32mDistributed Tracing in Action: Our Journey with OpenTelemetry
41mModern Application Debugging: An Introduction To OpenTelemetry
02h 02mHow To Troubleshoot and Monitor Applications using OpenTelemetry
01h 00mObservability made painless: Go, Otel & LGTM stack