Planet-Scale Dashboards

on YouTube ↗27m

" summary

Google builds a reusable dashboard system that automatically selects and filters dashboards for hundreds of thousands of services by detecting boolean traits-based discovery (eg. HTTP server, JVM) and applying scope types (server, database, ML model). It uses parameterized dashboards, metric label/variant layering and Prometheus-style recording rules to switch aggregation levels, cuts massive dashboard duplication and operational toil at Google scale, and runs in production across ~100 modeled scope types.

AI-written summary · about

" chapters (11)

1. 0:00Welcome to Planet-Scale Dashboards
2. 1:47Introduction
3. 3:55Traditional dashboard challenges
4. 6:24Parameterizing dimensions
5. 9:13Filtering by system traits
6. 13:02Scope types for resources
7. 15:54No infrastructure or code
8. 18:46Drilling down views
9. 21:13Decoupling metrics
10. 24:24Additional use cases
11. 25:47Key takeaways and closing remarks