The industry standard is to emit logs as [[JSON]]. Structured logging might be what you aim for when you don't know you should prioritize pure metrics with tools such as [[Prometheus]]. Actually, you should strive for both. Ship the logs to something like [[Humio]] or [[Splunk]] using a [[Docker]] sidecar container with [[Vector]] or similar. Also start thinking about using the [[AWS Distro for OpenTelemetry]] when working with [[Amazon Elastic Container Service|ECS]]. [The Twelve-Factor App from Heroku presents ideas on best practices for logging](https://12factor.net/logs). # Systematic review A short [[Systematic review]] (not really) on different logging solutions. ## Go Zerolog is recommended in [[Let's Go Further (book notes)]]. [A Complete Guide to Logging in Go with Zerolog | Better Stack Community](https://betterstack.com/community/guides/logging/zerolog/) is a great article. See [the discussion about structured logging in Go](https://github.com/golang/go/discussions/54763) for in depth comparisons. ## Python [structlog](https://www.structlog.org) is a popular choice. ## JavaScript Check out [Pino](https://portal.gitnation.org/contents/multithreaded-logging-with-pino) or [Winston](https://www.thedreaming.org/2020/06/24/structured-logging-nodejs/).