Diagnostics
Use developer diagnostics to understand setup readiness, active issues, passed checks and recommended fixes.
Diagnostics summarizes whether your Kadryn developer setup is production-ready.
Use it when Gateway, direct ingest, provider keys, webhooks, cost visibility or observability are not behaving as expected.
What Diagnostics checks
Diagnostics can evaluate signals across:
- authentication;
- provider keys;
- Gateway;
- direct ingestion;
- webhooks;
- costs;
- governance;
- observability.
It highlights critical, warning, info and passed checks.
Readiness states
| State | Meaning | | --- | --- | | ready | No blocking setup issue is detected. | | needs_attention | Some warnings should be reviewed. | | critical | At least one issue blocks or breaks a production path. | | waiting_for_signal | Kadryn needs traffic, delivery or ingestion data before it can judge readiness. |
Issue severities
| Severity | Meaning | | --- | --- | | critical | Fix before relying on this path in production. | | warning | Review soon; production may be degraded or incomplete. | | info | Useful guidance or context. | | passed | A check succeeded. |
Issue cards
Each issue explains:
- category;
- severity;
- title;
- impact;
- evidence;
- fix;
- last seen time;
- primary action;
- related docs.
Use the primary action to jump to the exact Kadryn page that can fix the issue.
Metrics
Diagnostics also summarizes operational signals such as:
- live usage events in the last 24 hours;
- Gateway requests in the last 24 hours;
- Gateway non-success count;
- fallback usage;
- direct ingestion attempts;
- rejected or invalid ingestion attempts;
- active webhook endpoints;
- failing or retrying webhook deliveries.
Recommended workflow
- Open Developers → Diagnostics.
- Review the readiness state.
- Fix critical issues first.
- Use the primary action on the top issue.
- Re-run a non-sensitive test.
- Verify results in Logs & Traces.
- Review passed checks.
When Diagnostics is locked
If Diagnostics is locked by plan or access, use the available developer pages directly:
- API Keys;
- Gateway;
- Direct ingest;
- Webhooks;
- Logs & Traces.
Upgrade or ask an admin if diagnostics are required for production operations.
Common issues
No live usage
Send a Gateway or direct ingest test request, then check Logs & Traces.
Gateway has errors
Open Gateway route readiness and provider keys. Check whether requests are blocked by policy or provider key errors.
Direct ingest rejected events
Open Logs & Traces and inspect validation errors. Check timestamp, provider, model, cost and metadata fields.
Webhook endpoint failing
Check response status, signature verification, endpoint latency and retry state.
Cost signals missing
Check whether usage events have pricing coverage and valid metadata.