healthcheck always fails due to connection refused

[opelx]

Describe the bug
heathcheck always fails due to connection refused

To Reproduce
Start the proxy and check

$ podman ps --format "{{.Image}}\t{{.Command}}\t{{.Status}}\t{{.Ports}}" |grep proxy
ghcr.io/wollomatic/socket-proxy:1               Up 11 minutes (unhealthy)       0.0.0.0:2375->2375/tcp
$ podman exec podman-socket-proxy ./healthcheck
2025/12/22 14:41:51 error doing health check: Head "http://localhost:55555/health": dial tcp [::1]:55555: connect: connection refused

For my understanding, ./healthcheck runs in the same network as the proxy self, sharing same user namespace.

I'm using rootless Quadlet with

[Container]
...
Environment=SP_ALLOWHEALTHCHECK=yes
Environment=SP_STOPONWATCHDOG=yes
HealthCmd=./healthcheck
...

Setting Environment=SP_STOPONWATCHDOG=no doesn't change this.

netstat doesn't shown anything, since the port 5555 isn't mapped to host.

Expected behavior
Should return true, otherwise false

Desktop (please complete the following information):

• OS: CoreOS stable/latest