Troubleshooting¶
Each page is a step-by-step diagnostic procedure for one symptom. Read the one that matches what you are seeing; do not skip steps — the order is calibrated so step N's evidence informs step N+1.
| Symptom | Page |
|---|---|
| Agent never appears in Fleet, or disconnects on first attempt | Agents do not connect |
Rollout sticks in InProgress or pauses unexpectedly | Rollouts stall |
| Operators cannot log in via OIDC | Login / OIDC issues |
| Database connection or migration errors | Database problems |
Information Ampora gives you for free¶
When something is broken, three places give you free signal:
/health/readyreturns a JSON body listing every probe. Failed probes have a free-text error.- The audit log records every state change with actor, timestamp, and before/after. Filtering by entity ("show me everything that happened to rollout
1234") usually grounds an investigation in 10 seconds. - Self-observability metrics for everything dispatch- and OpAMP-related.
If you have not wired self-observability up, Self-observability is the page to start with — debugging Ampora without metrics is artisanal.