Observability across serverless asynchronous managed services
summary
Explains how to stitch traces across Lambda plus SNS SQS EventBridge DynamoDB by instrumenting handoffs and data links, showing a working Medtronic & Lumigo backtrace that reconstructs async chains through queues and streams. Compares three concrete approaches: CloudWatch & Friends logs+metrics, home‑brewed trace propagation via message headers, and commercial serverless monitoring. Demos include locating pinpoint 1 invocation out of 1B, tracking deploy-induced latency changes, and measurable false‑positive rates.
AI-written summary · about
chapters (22)
- 0:00Intro
- 0:40Preamble
- 1:00Agenda
- 1:35Who am I?
- 2:19Serverless Changes Everything
- 2:32Serverless Components
- 3:46Serverless Impact
- 5:28Serverless Challenges
- 6:22Async Patterns
- 7:21Troubleshooting Serverless
- 7:47Severity
- 11:04Debug
- 12:52Solutions
- 13:02Option 1 Cloudwatch & Friends
- 14:12Option 2 Homebrewed Solutions
- 18:10Option 3 Serverless Monitoring Solutions
- 19:37Medtronic & Lumigo - Backtrace chain of async events
- 21:40Pinpoint 1 invocation out of 1B
- 22:06Optimize
- 22:33Track deploy effects
- 23:11Takeaways
- 24:00Thank you!
related talks
18mObservability in Serverless Applications
34mObservability in Serverless Application
46mObservability in an Asynchronous World
43mObservability for Serverless
49mMonitoring serverless applications on AWS: the monitoring monster becomes your friend.
39mHow to build observability into a serverless application
39mObservability in DevOps
59mModernize faster with serverless and end-to-end observability