Logging in the age of Microservices and the Cloud
summary
Advocates moving from SSHable files to centralized structured logs (JSON) with request IDs propagated in headers and MDC values copied across services, favoring an out‑of‑the‑box log host over DIY unless traffic exceeds extreme levels; compares in‑process appenders versus stdout-plus-agent, recommends decoupling storage from presentation, setting pragmatic retention, and standardize keys/values and timestamps for cross‑service queries. Cites CloudWatch Logs capacity of 15,000 messages per second per instance as an example of why buying often wins.
AI-written summary · about
chapters (15)
- 0:00Intro
- 1:30The good old days of logging
- 3:20The cloud
- 4:03Why are we logging
- 6:03Capturing log information
- 6:35Map diagnostic context
- 10:12Querying logs
- 12:52Cloud challenges
- 15:25Autoscaling
- 19:48Writing the logs
- 23:10Log retention
- 24:47Log messages
- 28:12Distributed system
- 28:35Trace request flow
- 31:39Query logs
related talks
10mStructured Logging and You
28mFrom Fragmented Logs to Unified Insights: Scaling O11y
29mAre logs a software engineer's best friend? Yes - follow these best practices
48mLogging and ranting
39mObservable microservices
20mFast log management for your apps
28mOperational visibility in distributed systems
29mCrucial Log Management Skills for DevSecOps