Service status and health

Where to look when the API misbehaves, and how to check it from your own monitoring.

Written By Carmine Cella

Last updated About 3 hours ago

Status page

status.datafuel.ai shows the current state of the API, the dashboard and the website, with incident history and 90-day availability. Subscribe there to get incident updates by email.

Health endpoints

Both endpoints are public and need no API key.

Endpoint	Answers
`GET /api/v1/healthz`	`200` while the API process is up.
`GET /api/v1/healthz?deep=1`	`200` while Postgres, Redis and the queue are reachable, `503` when any of them is not. Cached for five seconds.

Point your own uptime monitor at the deep check. It is the same probe the status page uses.

Capabilities

GET /api/v1/capabilities lists which task types and AI engines are switched on right now, with a reason for anything that is off. Operators switch a module off when an upstream provider is down or degraded. A call to a disabled module answers 503 with MODULE_UNAVAILABLE or ENGINE_UNAVAILABLE and charges nothing.

Check capabilities once before a big batch or an llm_scraping call, and when you see a 503 from the API.

Telling API trouble from target trouble

The API answers 5xx or times out on every request: check the status page, then the deep health check.
The API answers 200 but tasks come back failed or blocked: the target site is the problem. See Read the result envelope.
The API answers 503 with a code: a module is switched off. Check capabilities.

Reporting a problem

Contact support through the DataFuel dashboard with the task or job id, the timestamp and the code and message you received.