Structured Logging and You
summary
Advocates replacing printf-style messages with key/value fields so logs keep key values and systems can emit JSON to be parsed by machines. A concrete win: historical logs are easy to query and the team replayed six months of logs to simulate adding a cache and prove performance impact. Recommends JSON over ad hoc logfmt, storing logs in Postgres or column stores or tools like spunks or Cubana, and using the CLI tool human log to colorize and search locally.
AI-written summary · about
related talks
38mLogging in the age of Microservices and the Cloud
06mIgnite: Bulletproof Logging
20mFast log management for your apps
39mFast log management for your infrastructure
27mLogs Are Good, Actually
48mLogging and ranting
15mIs structured logging really the be-all and end-all?
36mBeyond grep: Practical Logging and Metrics