GraphQL Observability with Elastic and OpenTelemetry
summary
Demo shows a working pipeline that instruments GraphQL with OpenTelemetry, pushes OTLP traces directly into Elastic APM, and uses Hot Chocolate diagnostics plus a custom activity enricher to rename root activities for readable service maps. Observability uncovers the real problem: local 170ms responses hide 900ms production spikes and repeated SQL calls per asset caused by a malformed downstream API (batched HTTP became N calls). A complexity metric surfaces queries with weight 135 vs allowed 5, enabling alerts and targeted fixes.
AI-written summary · about
chapters (9)
- 0:00Introduction to observability
- 1:04GraphQL performance challenges
- 5:20Debugging in local environments
- 9:12Apollo tracing and limitations
- 15:24Implementing open telemetry
- 22:47Instrumenting your services
- 30:52Observability in production
- 34:19Refining telemetry data
- 45:09Advanced metrics and wrapping up
related talks
08mGraphQL Observability
23mGraphQL Observability
33mThe Observant Developer - OpenTelemetry from code to prod
53mUnmask hidden issues with Observability superpowers!
55mCloud-Native Observability with OpenTelemetry
37mObservability For You And Me With Opentelemetry
19mOops my app performance sucks... what if observability could be a game changer?
25mUnlocking Insights: Navigating the World of GraphQL Observability