Skip to main content
GET
Health check
The liveness and readiness probe for a Recurso instance. It pings Postgres, pings Redis when the process connected to one at boot, and reports the boot-time TigerBeetle connection state, then rolls the two pings up into a single status. No authentication is required, and the response never includes the underlying error — component failures are logged server-side only, so a public probe cannot leak connection details. Point your uptime monitor at this endpoint (see Monitoring and Alerting). For the build identifier alone, use Get Version; for scrape-style telemetry, use Get Prometheus Metrics.
The HTTP status follows the system of record. Postgres down returns 503 with status: "degraded". Redis down (locking and rate-limit degrade, money paths still work) returns 200 with status: "degraded". A disconnected TigerBeetle (the Postgres ledger is authoritative; TigerBeetle is an optional accelerator) does not touch status at all — with Postgres and Redis up you get 200 and status: "ok" alongside components.tigerbeetle.status: "disconnected". Read components for the detail.

Example Request

Response

Returned bare — no data envelope.
When Postgres is unreachable the same shape comes back with HTTP 503:

Fields

Errors

This endpoint does not use the error envelope — every response, including 503, is the health payload above.

Authorizations

Authorization
string
header
required

Response

Service is healthy (or degraded but serving).

status
enum<string>
Available options:
ok,
degraded
version
string
components
object