Health, Liveness & Readiness Probes

MTVL includes dedicated health probe endpoints designed for container orchestrators (Kubernetes, Docker Swarm, ECS, Fly.io).

Health Probe Summary

Endpoint	Probe Type	Checks Database	Expected Status
`GET /health`	General Health	Yes (Postgres/MySQL)	`200 OK` (Healthy) / `503` (DB Down)
`GET /healthz`	Kubernetes Health	Yes (Postgres/MySQL)	`200 OK` / `503`
`GET /livez`	Liveness Probe	No (Process only)	`200 OK`
`GET /readyz`	Readiness Probe	Yes (Postgres/MySQL)	`200 OK` / `503`
`GET /startupz`	Startup Probe	Yes (Postgres/MySQL)	`200 OK` / `503`

Checks if the HTTP server process is responsive. Does not ping the database to avoid cascading restarts during transient database network blips.

HTTP/1.1 200 OK
Content-Type: application/json

{"status":"alive"}

Executes a quick ping (SELECT 1) against the configured database.

If database responds: 200 OK -> {"status":"healthy","database":"connected"}
If database unreachable: 503 Service Unavailable -> {"status":"unhealthy","error":"database unreachable"}