> ## Documentation Index
> Fetch the complete documentation index at: https://docs.orinadus.com/llms.txt
> Use this file to discover all available pages before exploring further.

# GET /health

> Check intake server liveness and current journal state.

## Endpoint

```http theme={null}
GET /health HTTP/1.1
Host: 127.0.0.1:18799
```

No authentication required. No request body.

## Response — 200 OK

```json theme={null}
{
  "status": "ok",
  "events": 1042,
  "ephemeral": false
}
```

<ResponseField name="status" type="string">
  Always `"ok"` if the server is running and able to respond. There is no `"error"` state — the server either responds or it doesn't.
</ResponseField>

<ResponseField name="events" type="integer">
  Current event count from `Journal::stats()`. Returns `0` if the journal file does not exist or if `stats()` fails (e.g. permission error). Never returns an error for journal read failures.
</ResponseField>

<ResponseField name="ephemeral" type="boolean">
  Whether ephemeral (burn) mode is currently active. When `true`, `POST /ingest` requests return `202` and are discarded without writing to disk.
</ResponseField>

## Example

```bash theme={null}
curl -s http://127.0.0.1:18799/health
```

```json theme={null}
{"status":"ok","events":1042,"ephemeral":false}
```

## Use cases

* Liveness check in a health monitor or systemd unit
* Verify the server started before sending ingest requests
* Check whether ephemeral mode is active from outside the MCP session

## Notes

* The journal path is intentionally not included in the response. Use `urchin doctor` or `urchin_status` (MCP) to get it.
* The `events` count is read from the journal on every request. There is no cached counter.
* This endpoint does not require `urchin serve` to be running if you're calling it from the same process — but in practice, it is always served by `urchin serve`.