Exact meaning of HEALTHCHECK_PORT

[EugenMayer]

Before anything else, thank you for providing this awesome docker image!

Currently writing a help chart for this (which i will share) and asking myself, what the exact meaning of HEALTHCHECK_PORT will be.

I currently assume it will fail if backups fail, but maybe it's more a liveness probe? Could you explain under which circumumstanes it will be non 2xx?

Thank you a lot

[EugenMayer]

no need to keep questions open that are not going being be answered

[prodrigestivill]

Healthcheck is using the go-cron application healthcheck port.
I assume it will fail only on timeout when process is locked.

[motey]

Behind the port runs a webserver that outputs a report in json format. A healthy report looks like

{
  "Running": {},
  "Last": {
    "Exit_status": 0,
    "Stdout": "",
    "Stderr": "",
    "ExitTime": "",
    "Pid": 0,
    "StartingTime": ""
  },
  "Schedule": "* * * * *"
}

An unhealthy report can look like

{
  "Running": {},
  "Last": {
    "Exit_status": 1,
    "Stdout": "You need to set the POSTGRES_DB or POSTGRES_DB_FILE environment variable.\n",
    "Stderr": "",
    "ExitTime": "2023-11-02T12:27:00Z",
    "Pid": 42,
    "StartingTime": "2023-11-02T12:27:00Z"
  },
  "Schedule": "* * * * *"
}

I came up with a very ugly but working docker-compose healtcheck :)
It does check if there is anything running on the port. if yes it does parse the json with bash base tools and exit with the code that contained in Last.Exit_status

services:
  your_service:
    image: prodrigestivill/postgres-backup-local
    ports:
      - "8083:8083"
    environment:
      - HEALTHCHECK_PORT=8080
      - SCHEDULE=* * * * *
    healthcheck:
      test: ["CMD-SHELL", "curl -s http://localhost:8080 && exit $(curl -s http://localhost:8080 | grep -o '\"Exit_status\": [0-9]*' | awk -F: '{print $2}' | tr -d ' ')"]
      interval: 10s
      timeout: 5s
      retries: 3

I am pretty sure somebody can come up with a more elegant healthcheck :) but this works for now.
Another possible improvement would be to output the stdout and Stderr entries. Then a docker inspect or any monitoring tool that relies on docker healhtcheck messages would be able to deliver a meaningful error message.