Métrologie et Alerting avec Prometheus et Grafana

Prometheus est un outil open source qui permet de collecter facilement des métriques et de créer des alertes en fonction des valeurs récoltées. Venez découvrir comment mettre en place Prometheus et quelques collecteurs afin de monitorer l’état de santé de votre infrastructure. Je vous montrerais également qu’il est possible de construire vos propres métriques métiers avec SpringBoot afin de suivre la performance de vos applications au plus près et de les afficher sur un dashboard Grafana.

Speakers : Yoan Rousseau, Pierre-Yves Aillet

Transcript

1. Métrologie et Alerting Avec Prometheus et Grafana Pierre-Yves Aillet @pyaillet Yoan Rousseau @yoan_rousseau
2. Pierre-Yves Aillet Consultant formateur à Zenika Nantes @pyaillet @pyaillet Yoan Rousseau Consultant formateur à Zenika Rennes @yoan_rousseau @y0anR
4. Métrologie? Monitoring?
5. Le monitoring
6. Infrastructure
7. Middleware
8. Métier
9. Applicatif
10. Le monitoring
11. Le monitoring
12. Prometheus
13. Prometheus
14. CLOUD NATIVE COMPUTING FOUNDATION
15. • Multi-dimensional time serie ◦ A name ◦ Some labels ◦ A timestamp ◦ A float64 value • A flexible and easy query language: PromQL • Time series collection happens via a pull model Prometheus
16. • Raw log/ event collection (Es Stack) • Request Tracing (OpenTracing/Zipkin) • Durable Long-term Storage • Automatic horizontal Scaling • User/ auth mngt Prometheus
17. Prometheus
18. Prometheus
22. Prometheus
23. Prometheus
24. probe_http_duration_seconds{phase="resolve"} 0.020529207 Metric name Label Time series Prometheus - Metrics
25. Prometheus - Metrics Basic counters Sampling counters Counter Histogram Gauge Summary
26. Metric Timestamp Value http_request_total 1517585739 5124 http_request_total 1517595739 5101 http_request_total 1517605737 5200 Raw datas Storage
27. Metric Timestamp Value http_request_total 1517585739 5124 http_request_total 1517595739 5101 http_request_total 1517605737 5200 Raw datas Metric Timestamp Value http_request_total 1517585739 5124 http_request_total +10000 -23 http_request_total +19998 +76 Delta storage Storage
28. Metric Timestamp Value http_request_total 1517585739 5124 http_request_total 1517595739 5101 http_request_total 1517605737 5200 Raw datas Metric Timestamp Value http_request_total 1517585739 5124 http_request_total ++10000 +-23 http_request_total +-2 ++99 Double delta storage Metric Timestamp Value http_request_total 1517585739 5124 http_request_total +10000 -23 http_request_total +19998 +76 Delta storage Storage
29. Prometheus
30. Query language: http_requests_total http_requests_total{job="prometheus",group="canary"} http_requests_total{environment=~"staging|testing|development",method!="GET"} sum(http_requests_total{method="GET"} offset 5m) rate(http_requests_total[5m] offset 1w) -- https://prometheus.io/docs/prometheus/latest/querying/basics/ Prometheus QL
31. Query language: http_requests_total http_requests_total{job="prometheus",group="canary"} http_requests_total{environment=~"staging|testing|development",method!="GET"} sum(http_requests_total{method="GET"} offset 5m) rate(http_requests_total[5m] offset 1w) -- https://prometheus.io/docs/prometheus/latest/querying/basics/ Prometheus QL
32. Query language: http_requests_total http_requests_total{job="prometheus",group="canary"} http_requests_total{environment=~"staging|testing|development",method!="GET"} sum(http_requests_total{method="GET"} offset 5m) rate(http_requests_total[5m] offset 1w) -- https://prometheus.io/docs/prometheus/latest/querying/basics/ Prometheus QL
33. Query language: http_requests_total http_requests_total{job="prometheus",group="canary"} http_requests_total{environment=~"staging|testing|development",method!="GET"} sum(http_requests_total{method="GET"} offset 5m) rate(http_requests_total[5m] offset 1w) -- https://prometheus.io/docs/prometheus/latest/querying/basics/ Prometheus QL
34. Query language: http_requests_total http_requests_total{job="prometheus",group="canary"} http_requests_total{environment=~"staging|testing|development",method!="GET"} sum(http_requests_total{method="GET"} offset 5m) rate(http_requests_total[5m] offset 1w) -- https://prometheus.io/docs/prometheus/latest/querying/basics/ Prometheus QL
35. Prometheus
37. Prometheus
38. Prometheus UI
39. Grafana
40. Demos
41. Feedback
42. • Server-side metric metadata support • Prometheus metrics format as a standard • Backfill time series • Support the Ecosystem Prometheus - Roadmap
43. Trugarez ! Questions?
44. RÉFÉRENCES https://prometheus.io/docs/instrumenting/exporters/ https://www.robustperception.io/scaling-and-federating-prometheus/ https://kjanshair.github.io/2018/02/20/prometheus-monitoring/