Giraffe: our journey to support 1 million metrics per second

Project Giraffe
our struggle journey to support 1 million metrics per second

Transcript

1. @sebamontini MEDALLIA Project Giraffe our struggle journey to support 1 million metrics per second
2. @sebamontini MEDALLIA Sebastian Montini SRE Sr. Manager @ Medallia https://speakerdeck.com/sebamontini https://linkedin.com/in/sebamontini
3. @sebamontini MEDALLIA Why?
4. @sebamontini MEDALLIA Observability
5. @sebamontini MEDALLIA A system is observable if you can determine the behavior of the system based on it’s outputs.
6. @sebamontini MEDALLIA A system is observable if you can determine the behavior of the system based on it’s outputs.
7. @sebamontini MEDALLIA A system is a set of connected components.
8. @sebamontini MEDALLIA A system is observable if you can determine the behavior of the system based on it’s outputs.
9. @sebamontini MEDALLIA The manner in which a system acts is it’s behavior.
10. @sebamontini MEDALLIA A system is observable if you can determine the behavior of the system based on it’s outputs.
11. @sebamontini MEDALLIA The outputs of a system are the concrete results of it’s behaviors.
12. @sebamontini MEDALLIA Monitoring is the action of observing and checking the behavior and outputs of a system and it’s components over time.
13. @sebamontini MEDALLIA The big 5
14. @sebamontini MEDALLIA Instrumentation Collection Storage Alerting Visualization
15. @sebamontini MEDALLIA
16. @sebamontini MEDALLIA The Four Golden Signals
17. @sebamontini MEDALLIA Latency Traffic Errors Saturation
18. @sebamontini MEDALLIA How?
19. @sebamontini MEDALLIA Giraffe
20. @sebamontini MEDALLIA “Giraffe is a scalable metrics platform, designed with reliability and scalability in mind. Giraffe is based in Prometheus and Thanos, running on top of Kubernetes (EKS) within Amazon Web Services” What is it?
21. @sebamontini MEDALLIA • Kubernetes (EKS): ”We run all the components within a K8S cluster. This give us the ability to be fault tolerant, more dynamic with resources, rescheduling capabilities, service discovery, easy deployments, auto scaling and much more.” • Prometheus: “Prometheus is an open-source project for monitoring and alerting originally built at SoundCloud. This project has two main components (tsdb and alert manager) that give us the power to manage data.“ • Thanos: ”Thanos is an open source project by Improbable.io to seamlessly transform existing Prometheus deployments in clusters around the world into a unified monitoring system with unbounded historical data storage.” • Telegraf: ”Telegraf is being used to ship metrics from DC’s to Giraffe without any change of the current deployment.” Main Components
22. @sebamontini MEDALLIA Prometheus
23. @sebamontini MEDALLIA Thanos
24. @sebamontini MEDALLIA Thanos
25. @sebamontini MEDALLIA
26. @sebamontini MEDALLIA Thanks