Supercharging API Monitoring with Guzzle Middleware - Liam Hammett
summary
Claims an operational system handling millions of API calls per week built around Guzzle middleware that captures structured logs to Kibana, errors to Sentry, traces to OpenZipkin/Jaeger (via Tempo) and metrics as Prometheus metrics with Grafana, Alertmanager and KEDA. Shows middleware code for request/response logging, OpenTelemetry spans and counters, Grafana dashboards and alert rules. Demo uses a Prometheus scrape workaround because the OTLP metrics exporter fails in the demo, and the repository and setup scripts are published.
AI-written summary · about
chapters (13)
- 0:00API monitoring with Guzzle
- 6:25Observability and telemetry
- 8:55OpenTelemetry standards
- 13:55OpenTelemetry in PHP
- 15:40Guzzle middleware
- 19:08Logging API requests
- 23:43Logs with Loki and Grafana
- 25:29Tracing API calls
- 28:59Traces in Tempo and Grafana
- 32:52Metrics with Prometheus
- 41:38Grafana dashboards
- 44:46Alerting in Grafana
- 48:20Resources and closing
related talks
33mBad API Observability
28mImprove Monitoring and Observability for Kubernetes with OSS Tools
02h 03mWorkshop 3
50mApplication metrics with Prometheus
56mFull-stack observability and application monitoring with
37mLowering the barrier for ML monitoring
52mGraphQL Observability with Elastic and OpenTelemetry
53mUnmask hidden issues with Observability superpowers!