Logging
Structured logs, log levels, and where they're shipped — what gets recorded and why.
Why we need this / what value this brings
Structured, well-placed logs are often the fastest way to actually diagnose a production issue.
When to use this
Every service boundary and every error path, at minimum.
How to use or implement this
Use structured (JSON) logs with consistent fields (request ID, user ID where relevant) so they're queryable, not just readable.
Empty folder — drop notes, links, and findings here as you research.