Status and health
The public status page, and the runtime health endpoint your monitoring can read.
status.intertrace.ai
The public status page shows each component, 90 days of history and incidents. Subscribe by email, or follow the RSS feed (/status/feed.xml) or calendar (/status/calendar.ics); a status badge is at /status/badge.svg.
| Group | Components |
|---|---|
| Control plane | Dashboard, Sign-in, Dashboard data, Background jobs, Platform schema |
| Data plane | Telemetry ingest, AI runtime gateway (LLM, MCP and authorize) |
| Detection | Optional sidecar classifiers |
Background jobs (Guardian reviews, session close, the activity ledger, SOC SLA escalations, runtime findings, assessments) read degraded when one has not completed within about twice its interval.
GET /v1/runtime/health
Call it on the gateway with a runtime key. status is ok, or degraded when any block below is failing.
curl -s https://intertrace.fly.dev/v1/runtime/health -H "Authorization: Bearer itr_rt_YOUR_RUNTIME_KEY"| Block | What it tells you |
|---|---|
agent_context | The policy context the enforcer is using: state, snapshot status, versions, last success and failure, expiry. degraded is true while it serves the last known good context or fails closed. |
judge | The AI judge: configured, provider, model, status (ok, unknown, failing), consecutive failures, last error (for example http_401:invalid_api_key) and when. |
guard_model | The local prompt-injection models: status, model, consecutive failures, last error. |
pii | The personal-data detection service: status and last error. |