Runtime logs

Runtime logs record what the service itself is doing, not your requests — startup and shutdown, listen-port changes, config saves, route version publishing, and internal errors and warnings.

Difference from request logs

	Runtime logs	Request logs
Subject	The service itself	The user requests that pass through
Question	Is the service working correctly	Why did this request fail
Typical content	Startup, config, internal errors	Dispatching, attempts, latency, upstream errors

The two logs are not mixed

Runtime logs serve troubleshooting — fields are request ID, latency, error code. Telemetry events serve product decisions and must be anonymous and low-cardinality. The requirements of the two are exactly opposite, so they are kept separate.

Common uses

Startup failure: port in use, data directory not writable.
Config anomalies: cloud-sync conflict, key decryption failure.
Routing issues: whether a particular publish succeeded, and whether there was a rollback.

Retention

Runtime log retention is tuned in Settings → Data → Log retention. It is a separate setting from the request log retention policy.

Next, see where data lands and how it is retained and cleaned — Data directory & retention.