OPERATIONS GUIDE
Monitor API data reliability
Zorqelio supports two monitoring modes in the dashboard REST API: push-based freshness monitoring and scheduled polling of a public HTTPS JSON endpoint. Both validate data against the contract’s latest published schema.
Push monitoring (recommended for private APIs)
Your own service, workflow or scheduled job sends each upstream JSON result to Zorqelio. This works when your API is private or requires authentication because Zorqelio never needs to fetch the source endpoint.
- Create a contract and note its ID.
- POST to
/api/v1/monitorswith the workspace token, contract ID, label and expected interval. - Store the returned monitor bearer token in your server-side secret manager.
- After each successful upstream fetch, POST the JSON response body to the monitor check endpoint.
- Review decisions and freshness in the dashboard or MCP
monitor_status/monitor_historytools.
curl -X POST https://zorqelio.xyz/api/v1/monitors/YOUR_MONITOR_ID/check \
-H "Authorization: Bearer $ZORQELIO_MONITOR_TOKEN" \
-H "Content-Type: application/json" \
-d '{"id":"evt_42","state":"ready"}'
A valid payload returns HTTP 200 and ALLOW. A contract violation returns HTTP 422 and BLOCK with issue evidence. A revoked monitor returns HTTP 410. The body is validated in memory and is not retained.
Freshness states
- healthy — a validly authenticated submission arrived within the expected interval.
- stale — the expected interval elapsed without a fresh submission; a five-minute Cron Trigger evaluates overdue monitors.
- revoked — the monitor token has been revoked and submissions are no longer accepted.
Expected intervals are 5–1,440 minutes; default is 60. A new authenticated submission clears stale state. Freshness is a heartbeat signal: it does not prove that the upstream API is correct unless its response is submitted and validated.
Scheduled public API checks
For a public JSON API that needs no authentication, create a monitor with an optional endpointUrl field in the dashboard or REST API. Example:
curl -X POST https://zorqelio.xyz/api/v1/monitors \
-H "Authorization: Bearer $ZORQELIO_WORKSPACE_TOKEN" \
-H "Content-Type: application/json" \
-d '{"contractId":"YOUR_CONTRACT_UUID","label":"Public status API","expectedIntervalMinutes":15,"endpointUrl":"https://status.example.org/api/status"}'
Zorqelio performs an HTTP GET at the configured interval, parses JSON and validates it against the latest contract version. The dashboard displays scheduled mode, HTTP status, latest error and check counts.
- Public HTTPS endpoint on port 443 only.
- No credentials, query strings, fragments, IP literals, local/private hostnames or own Zorqelio domain.
- Redirects are not followed.
- No authentication headers or custom request headers.
- Response must declare
application/jsonor a+jsoncontent type. - Response body limit 256 KB; request timeout 8 seconds.
- Up to 10 due scheduled monitors are processed per five-minute Cron run.
Events and notifications
Monitor history includes creation, stale transitions, recovery, blocked checks and revocation. The dashboard can show browser notifications while open and permission is granted. Email delivery is not active until a production sender is configured; no email alerts should be expected at this time. Slack, SMS and outbound customer webhooks are not currently provided.
Security and retention
Monitor tokens are shown once and stored as SHA-256 hashes. Rotate or revoke exposed credentials using the available management tools. Submitted payloads are not stored; bounded validation summaries and monitor events are retained for operational history. Do not place tokens in browser code, source control or public logs.
Incident lifecycle
When a monitor becomes stale or a response is blocked, Zorqelio creates one active incident for that monitor and failure category. The incident includes severity, a short explanation, and a recommended next step. Workspace members can acknowledge it or mark it resolved after manual investigation. When the monitor later reports a healthy recovery, active incidents for that monitor are automatically resolved with an automatic-recovery audit marker. This release does not yet retry upstream calls, quarantine payloads, or make changes to customer systems.