A Better Way to Manage Stateful Systems: Design for Observability and Robust
summary
Built a business-critical WebSocket connect server that holds long‑living tunnels and proxy connections, discovers blue‑green fails because sessions survive cutover, and adds an API to expose internal state that returns counts and metadata for sessions and test connections. A deployment pipeline using AWS Step Functions and Lambda checks metrics, marks idle sessions, invokes an endpoint to disconnect sessions, then drains and terminates nodes. Logging and tracing use session ID in OpenTelemetry spans so searches tie logs/traces to the exact state.
AI-written summary · about
chapters (5)
- 0:00Start
- 3:46A failure story-with a Blue-Green deployment
- 6:39Design tips for making stateful apps observable (2)
- 10:16st step
- 10:56th step
related talks
24mObservability across serverless asynchronous managed services
14mState of Serverless DevEx & Observability
49mImproving Observability of Distributed Systems
49mMonitoring serverless applications on AWS: the monitoring monster becomes your friend.
46mObservability in an Asynchronous World
34mObservability in Serverless Application
39mHow to build observability into a serverless application
43mObservability for Serverless