How I Solved Production Issues with OpenTelemetry (And How You Can Too)
summary
Uses OpenTelemetry Java agent and the collector plus Grafana/Loki to debug real production failures without changing code: discovers containers growing to ~8 GB because JVM max heap is unrestricted and fixes instability by capping containers to 250 MB; correlates traces and logs to find a synchronized block that causes P95 latencies up to 10s; instruments MS Graph calls, counts 15 SharePoint fetches, adds caching to eliminate dozens of extra HTTP calls; generates a service graph from traces; runtime overhead is about 1–2%.
AI-written summary · about
chapters (10)
- 0:00Introduction to observability
- 2:06Understanding OpenTelemetry
- 5:24OpenTelemetry architecture
- 10:59Demonstrating production fixes
- 16:09Correlation and root cause
- 24:57Analyzing latency and bottlenecks
- 30:12Advanced observability patterns
- 35:37Service graph visualization
- 38:07Summary and code challenges
- 42:36Q&A session
related talks
31mTroubleshooting and Monitoring Production Applications using OpenTelemetry
49mAdding OpenTelemetry to Production Apps Lessons Learned
31mUnlocking the Power of OpenTelemetry: Enhancing Design, Development...
17mOpenTelemetry Looks Good To Me: Loki, Grafana, Tempo, Mimir
01h 02mWhat is your app really doing? Make it tell you with OpenTelemetry.
35mOpenTelemetry for Busy People
49mDistributed Tracing in Java? OpenTelemetry for the Rescue!
47mUsing metric and traces to diagnose our cloud-native applications