Building for the Unknown: Scaling Observability as a Product
summary
Claims that treating observability as a product and applying platform engineering turns telemetry into consistent, self‑service insight for a portfolio of 85 products, ~14,000 customers and millions of users (he cites 1.5M and 3.5M figures) across 100 engineering teams. Builds an observability platform that centralizes Honeycomb/Datadog/Dynatrace telemetry, enforces hashing emails via pattern recognition for compliance, and ships SLO-driven alerts such as 200 ms with a 98% target so alerts are meaningful rather than noisy, which reduces team mistakes and scales visibility.
AI-written summary · about
chapters (12)
- 0:00Introduction to topic
- 1:08Software engineering basics
- 3:13Feature lifecycle in production
- 4:16Telemetry vs observability
- 6:23Scale of user impact
- 7:26Silent users problem
- 9:11Observability as a product
- 10:56Platform engineering approach
- 12:23Key takeaways summary
- 14:10Q&A: Alerting goals
- 16:20Q&A: Observability customers
- 18:07Q&A: Product mindset
