GraphQL Observability
summary
Shows how resolver-level traces from Apollo Server get shipped to observability backends via a vendor-agnostic spec, capturing p95/p99 metrics and per-resolver spans so an alerts workflow can pinpoint hotspots — e.g., notifications aggregation drives requests toward ~2s, an orders query totals ~700ms. Uses Apollo Studio for schema change and operation analytics, exports traces to OpenTelemetry collectors and visualizes in Honeycomb Apollo tracing or Jaeger/Datadog via a simple Apollo Server plugin; demo includes docker-compose usable on Kubernetes and notes ~20% sampling tradeoffs.
AI-written summary · about
chapters (5)
related talks
08mGraphQL Observability
52mGraphQL Observability with Elastic and OpenTelemetry
25mUnlocking Insights: Navigating the World of GraphQL Observability
17mUnlocking Observability with React & Node.js
26mBeyond HTTP 200: Observability With GraphQL
56mFull-stack observability and application monitoring with
45mObservability is for User Happiness
33mBad API Observability