Health
2 endpoint(s)
GET /health
Liveness probe
Process is up. Independent of upstream dependencies — K8s should not restart this container just because Fracttal DB is degraded.
@limiter.exempt is critical: the kubelet probes /health from the cluster
network without an Authorization header, so it falls into the IP bucket.
Under load the kubelet's source IP can saturate that bucket → 429 → the
kubelet treats it as a liveness failure → container restart cascade.
Found via FHEI-1121 capacity test (1.1.26-test): 4 self-induced restarts
in 20 minutes. Same rationale for /ready.
Respuestas
| Código | Descripción |
|---|---|
200 |
Successful Response — HealthResponse |
GET /ready
Readiness probe
Both Fracttal backends (PG, RPC) must report healthy circuit breakers.
Autenticación
Bearer Token requerido. El id_company se extrae automáticamente del token.
Respuestas
| Código | Descripción |
|---|---|
200 |
Backends healthy — accepting traffic |
503 |
At least one backend breaker is open — drain from Service |