How to build observability into a serverless application
summary
Shows how to implement log aggregation, tracing and correlation IDs for serverless apps using CloudWatch Logs, API Gateway, SNS/SQS/Kinesis and X‑Ray; recommends stream CloudWatch Logs to Kinesis so log-shipping uses shard-to-concurrency mapping (avoids hitting the regional 1,000 concurrency soft limit), uses structured JSON logs, and encodes custom metrics as log lines to avoid user-facing latency. Implements a middleware that propagates request IDs and sample 1% debug logs across sync and async invocations and stores correlation IDs in a per-invocation global so traces across functions can be rebuilt.
AI-written summary · about
chapters (5)
- 0:00Start
- 11:48building event-driven architectures
- 21:40enable sample logging in our application
- 25:00use the media middleware
- 34:03set up metric filters
related talks
34mObservability in Serverless Application
24mObservability across serverless asynchronous managed services
43mObservability for Serverless
41mLive Coding on Observability for Serverless
32mBuilding and scaling a log analytics platform - a serverless approach
40mCloud architecture patterns, how to implement them in a serverless way
49mMonitoring serverless applications on AWS: the monitoring monster becomes your friend.
51mServerless architectural patterns