GET
/health
Service status, including label-cache age and backing stores.
Reports whether each backing store answered and how stale the in-process label cache is. The interesting field is labels.age_sec: the service stays up and answers with a cache older than it should be rather than failing closed, so a monitor that only checks HTTP 200 will not notice staleness.
Request
curl https://intelapi.publicaml.org/healthResponse
{
"build": "6a7d4ecc8947",
"pg": true,
"redis": true,
"rocksdb": true,
"labels": {
"ok": true,
"rows": 295840,
"age_sec": 340,
"chains": { "0": 79572, "1": 149191, "56": 57163, "195": 9914 }
}
}Response fields
| Field | Type | Meaning |
|---|---|---|
| build | string | Hash of the running code. |
| labels.age_sec | number | Seconds since the label cache was last loaded. The freshness signal. |
| labels.chains | object | Loaded label rows per chain id. |
| pg / redis / rocksdb | boolean | Backing store reachability. |
Things that catch people out
- Alert on labels.age_sec, not only on the status code. A 200 with a stale cache is the failure mode this endpoint exists to expose.
Try it
The interactive console runs this endpoint against live data from the browser, with no key and no signup.
Open the sandbox at intelapi.publicaml.org