Defining Service Level Objectives for Exposure Notifications Servers
summary
Defines an availability SLO initially set to 99.5% SLO using the Google Cloud Load Balancer https request count metric, then narrows scope to the user-facing API server only after noisy alerts and misclassified HTTP codes flood on‑call. Moves SLO and alert config into Terraform so changes get code review, swaps latency SLOs for threshold alerts where interactions are asynchronous, and cites real impact: 1% adoption → 0.8–2.3% fewer cases.
AI-written summary · about
chapters (7)
- 0:00Service level objectives intro
- 3:34Reliability matters
- 7:47Steps to define SLOs
- 8:29Exposure Notifications overview
- 13:50Key SLO lessons
- 15:58Reliability impact
- 16:40Conclusion
related talks
10mExposure Notifications: Building infrastructure to serve 1M users
42mGuaranteed SLAs with higher-order functions (...)
35mFrom Monitoring to Observability: Left Shift your SLOs with Chaos
23mScalable security: how to win friends and not burn...
37mLevel Up Your Edge: Do You Know Where Your Traffic Comes From?
41mSoftware Delivery Shield: securing the serverless software supply chain
25mReliability To the Edge: Telco Teaching Cloud Native
33mHow to implement DDoS protection in GCP