Linux SWAG container - Is there an option for a healthcheck?

General Docker
1

Is there a method to configure via docker-compose a healthcheck on the swag container? I’m aware to configure a healthcheck via a Dockerfile but this would be an add on – taking the base swag container and adding a healthcheck. I didn’t know how if this capability was already built in.

2

Yes, you can define a healthcheck in your compose: Docker Tip #85: Define HEALTHCHECK in your Docker Compose File — Nick Janetakis

3

Hey thanks for reply

I’m aware of using curl or other such tools like wget to healthtest a container, however usually there is a port or endpoint that will answer back OK or something similar. In the example you linked to the command:

test: "${DOCKER_HEALTHCHECK_TEST:-curl localhost:8000/healthy}"

localhost:8000/healthy probably answers back OK or something similar.

I was wondering if there was a similar endpoint in the swag container which I could query for a healthcheck.

4

SWAG is an nginx webserver, so just point it to any address/port that curl can successfully connect to (like the homepage) or just have nginx serve a 200 at an address you pick.

1 Like
5

@aptalca

Thanks for the suggestion, after just a little bit of reading, I made everything work as you hinted. I know this is such an easy fix, but in general it was kind of hard as a beginner to see how everything came together to make a healthcheck work successfully. The official docker documentation discusses healthcheck and its parameters but never really discussed how to implement such a healthcheck with a container – for example in the use of this nginx container it meant declaring a location block like the following within my proxy subdomain:

server {
        server_name disco.domain.com;
        listen 8000;

        location /health-check {
                add_header Content-Type text/plain;
                return 200 "OK\n";
        }
}

The healthcheck then in my docker-compose file would be implemented like the following:

    ports:
      - 443:443
      - 80:80
      - 8000:8000
    restart: unless-stopped
    healthcheck:
      test: "${DOCKER_HEALTHCHECK_TEST:-curl localhost:8000/health-check}"
      interval: "60s"
      timeout: "3s"
      start_period: "5s"
      retries: 3

After figuring this out, I’m aware healthchecks could be implemented in like a hundred different ways, however just weird I didn’t find any specific basic examples like this putting it all together.

Thanks for your help!!!

6

Also keep in mind that if there is an available page that loads, you can just use that (ie. no auth). I don’t believe it needs to have the content “ok”. It just needs to load properly with a 200 return code (successful connection).

7

@aptalca – Thanks for information. Definitely I could simply the setup in most cases with your information. In this particular setup I’m using nginx in front of a reverse proxy for syncthing’s discovery server. I’m not sure exactly how the application exactly works, but it does not serve or return web pages. I’ve read technically to return a 200 something needs to be returned whereas serving no content is technically a 204 – which in some instances breaks a healthcheck in some implementations.

8

I meant you can use any existing page that loads without auth. For instance if you want to add a healthcheck to tautulli, you can use tautulli’s homepage. You don’t need to create/serve a new page just to satisfy the healthcheck. If swag’s home page loads without auth (perhaps on local networks, which you can do with allow/deny statements in the site conf) then you can just use curl -kL localhost:80