Webhooks

When I set up my new (well now newish) server, I gave myself the challenge of running as much as possible in docker containers. I wasn’t able to satisfactorily run Hugo itself in a container, but I don’t mind much since it’s a static site generator and I think Docker would be largely pointless. I wanted to try and get webhooks working, also, but wasn’t able to in Docker either.

I am certain there’s more thinking and research I need to do; it must be possible to manage webhooks in a docker container, but since we rely on running git and hugo to build the site, I can’t think of a way to do it short of creating a new image. I could do that, but I just feel it would be needless complexity for something of this scale. At the very least, I’ve figured out the basics of webhooks and that’s fine, and as a result I can push directly to my repository on GitLab and instantly have posts and content updates here be published, without needing to log in to my server.

I’ve achieved this using adnanh’s webhook server and I keep it running on the host, managed with supervisor. It’s not a great solution when you’re trying to keep things in containers, but it’s fine. My script for pulling and rebuilding the site is also highly rudimentary; just two lines: git pull and hugo -D. Not the best, and contains no failsafes, but it’s fine for now. I also don’t love that the webhook server is exposed on the host, but some access rules could alleviate any possible grief from this, and since GitLab will send requests over HTTPS with a secret token, it’s not so bad.

All in all, I’m happy with the new set up. It’s taught me what I need to know about webhooks and I’ve devised a way to author a website statelessly. Pretty neat.