Building scalable OSS observability with Mimir, Loki, Tempo, and Pyroscope
summary
Covers concrete changes and measured results for Grafana Mimir, Loki, Tempo, and Pyroscope: Mimir adds native histograms and exemplars to link outlier metrics to traces; Loki 3.0 builds query acceleration with n‑gram Bloom filters that cut chunks scanned by ~72–93% in tests (near‑100% for UUID filters); Tempo adds TQL metric summaries and structural operators to aggregate by endpoint and user ID; Pyroscope offers per‑trace profiles and flame‑graph diffs (1–3% overhead) that pinpoint a function jumping from 31% to 75% CPU.
AI-written summary · about
chapters (11)
- 0:00Introduction: Hello world!
- 2:10The cost of siloed observability at scale
- 3:30The "true" definition of observability
- 5:00Using the full observability toolkit
- 6:07How to find anomalies in metrics trends
- 9:33The history of Loki
- 11:38Loki 3.0 is coming - here's what to expect
- 15:07Query acceleration (flagship feature of Loki 3.0)
- 20:45How to find where latency and errors originate across many services with traces
- 29:55How to find root causes of issues at the code level
- 37:20Profiling: Debugging an incident (Demo)
related talks
01h 02mThe Power of Metrics, Logs & Traces with Open Source
19mBest Practices for Rapid Insights from Metrics, Logs, Traces and Profiles
32mWhat’s new with Grafana Labs’s Open Source Observability stack
17mOpenTelemetry Looks Good To Me: Loki, Grafana, Tempo, Mimir
27mObservability with Prometheus and beyond
36mObservability with Prometheus and Beyond: Leveraging Cloud-native Technologies Outside of Cloud
17mFrom noise to signal: piping your way to clear observability
53mUnmask hidden issues with Observability superpowers!