M3 and Prometheus, Monitoring at Planet Scale for Everyone
summary
M3 is an open source metrics platform that integrates with Prometheus via prometheus remote_write and runs on Kubernetes with an operator to deploy M3Coordinator and M3DB; it stores petabytes, indexes billions with FST segments, and replicates across AZs for HA. Real deployments at Uber see ~700M aggregated metrics peak, ~35M stored, ~9B total metrics, and demo ingestion reaches millions of writes/sec; it performs real-time downsampling and supports retention tiers like 10s→2d, 1m→30d, 1h→5y.
AI-written summary · about
chapters (10)
- 0:00Introduction
- 0:40Overview
- 4:53M3 Overview
- 9:07M3 Runs Anywhere
- 9:32M3 with Prometheus
- 13:06Graphite
- 14:16M3 Deployment
- 23:03Automation
- 25:10Demo
- 33:10QA
related talks
31mLong-Term Metrics with M3 and Prometheus
41mLearnings, patterns and Uber’s metrics platform M3...
11mAggregating, Alerting and Graphing on Millions of Prometheus T...
39mHow To Take Prometheus Planet Scale: Massively Large Scale Metrics Installations
26mWhen Prometheus Can’t Take the Load Anymore
42mPrometheus Intro and Deep Dive
10mPrometheus for Monitoring Metrics - Database Monitoring Tools
31mWhen Prometheus Can’t Take the Load Anymore