Publish Docker images to GHCR or Dockerhub for self-hosting

[Blankeos]

The repo already has Dockerfiles (api.Dockerfile, basket.Dockerfile, links.Dockerfile, uptime.Dockerfile) but no pre-built images are published to any container registry.
This blocks one-click self-hosting on platforms like Dokploy, Coolify, and CapRover — all of which rely on pulling pre-built images rather than building from source.
Proposal: Add a GitHub Actions workflow that builds and pushes images to GHCR on each release, e.g.:

• ghcr.io/databuddy-analytics/databuddy-api:latest
• ghcr.io/databuddy-analytics/databuddy-basket:latest
• ghcr.io/databuddy-analytics/databuddy-links:latest
• ghcr.io/databuddy-analytics/databuddy-uptime:latest

Tagged with both latest and the version (e.g. v1.0.0). Those would live here.
For reference, Rybbit (https://github.com/orgs/rybbit-io/packages) does exactly this and has a working Dokploy template + (Github source) as a result.

[sekhar08]

Is someone working on this issue?

[izadoesdev]

Working on it now

[sekhar08]

@izadoesdev can i work on it?

[izadoesdev]

@izadoesdev can i work on it?

Sure, I've already created a basic implementation but if you wanna improve & get it more ready for self-hosting, that'd be great

[sekhar08]

@izadoesdev, Should I also create a Dokploy template, or is the GHCR publishing workflow sufficient for now?

[izadoesdev]

Self hosting and templates are still a bit far fetched for now, but this would be a great first step

I'm saying this because a lot of things are hardcoded to work in Databuddy's environment, and there's a lot of things that are overkill for self-hosting (Redpanda, maybe even clickhouse can be optional)

But it's definitely planned if you wanna get started on it 👀

[sekhar08]

Got it! I'll focus on just the GitHub Actions workflow for now to get the GHCR publishing set up. Once that's merged and working, we can tackle the Dokploy template after the hardcoded dependencies are refactored for self-hosting.

how does that sound?

[izadoesdev]

Sounds good, ping me once the first step is done, then we can set up a plan for self-hosting

[sekhar08]

Hey @izadoesdev, can you take a look at #371 and suggest changes if needed? Thanks in advance.

[sekhar08]

Hey @izadoesdev, I've asked clude for a plan for self-hosting, it gave me an md file regarding it. Here is a gist of what it gave in https://gist.github.com/sekhar08/67d0df56091e6430aa85ad1768bde2ac . would like to hear your thoughts on this.

[izadoesdev]

Looks like a good start, but there's a lot of things to consider, like ideally we want databuddy to be lightweight, which means redpanda and clickhouse might not be needed for 99% of users

• Make all URLs configurable via env vars
• Create production docker-compose.yml
• Add health checks to all Dockerfiles

You can work on these if you want, I'll tackle the upstash dependancy, either via BullMQ or just skipping it

[sekhar08]

Hey @izadoesdev, I've worked on the issues mentioned. can you please look at the PRs #374, #375 and #379. Thank you in advance.

[sekhar08]

Hey @izadoesdev, really appreciate you for making this project. do you plan to release more issues anytime soon? I'll be glad to contribute to Databuddy more meaningfully. Thank you :)