Observability

How observability works in Galaxy

1. 1.

   Health & events — a live, complete record.

   Service health and metrics surface degradations as they happen, while every query, export, permission change, and sync writes an immutable event with the actor, the target, and a timestamp. Nothing meaningful happens off the record.

2. 2.

   Traces — follow a request end to end.

   A trace stitches the related spans of a single action together, so you can see exactly what a query touched, which services it reached, and where the time went.

3. 3.

   Alerts — know when something's off.

   Define rules — a failing sync, latency over threshold, schema drift, an external user exporting restricted data — and Galaxy raises an alert the moment the pattern appears.

4. 4.

   Lineage — where data came from.

   Lineage maps how a value flowed from source to result, so when a number in a report looks wrong you can walk it back to its origin.