4. Logging That Finds You

Coming soon

This chapter is an outline. The full draft ships in a follow-up PR.

What this chapter will cover

Logs in a Uni application are written for a human reading output, not for a log-aggregation pipeline to parse. That framing shapes what Uni's logger does, and — just as importantly — what it chooses not to do.

Concepts introduced:

• The LogSupport trait and why mixing it in beats passing a Logger around.
• Source-location capture: how Uni captures file and line at compile time, and what that means at runtime.
• Log levels and how to choose between them without lying to your future self.
• Per-package log configuration at runtime.
• Why Uni's logger intentionally has no structured-logging API: logs are developer context, not machine-parsed records. When you need fields, IDs, and correlation, reach for a telemetry tool (metrics, traces) instead of encoding state into log lines.
• Formatting choices for terminals and for log files.

Reference you can read now

• Logging (module reference)

← 3. Wiring with Design | Next → 5. JSON & MessagePack