Can we host dagger module in artifactory?

[vagrant phoenix]

Hello Team 👋

Is possible to host dagger module other than vcs providers(gitlab, gtihub), lets say in jfrog artifcatory?

we use gitlab saas and we are seeing errors during module download by dagger engine due to gitlab api call limit(shared gloablly 5 per second or something), even though caching module is saving us for repetitive calls but when cache invalidates, it sometimes get into the trouble .

I am looking for a way improve this 🙂

[unique shale]

cc @fleet turret

[unique shale]

I think it will be ok but I don't think we have tested it. Let us know if you run into any issues

[vagrant phoenix]

@unique shale I am not sure if that should be released as a zip artifact or code file directly, any thoughts?

[unique shale]

I thought artifactory also had some sort of git support. Modules can only be loaded from git and local sources currently

[fleet turret]

Hey Akki,

As Marcos said, Dagger can only pull modules from Git or local paths. Standard JFrog Artifactory can’t serve as a Git host, so to get around GitLab’s ~5 req/sec API limit you’ve got three workable routes:

1. Internal Git mirror – Mirror the repo to Gitea/Gitolite or a bare SSH Git server.

How: Dagger fetches from the mirror, avoiding GitLab limits.
Pros: Same Git workflow, full history.
Cons: Needs infra + sync from GitLab.

2. Pre-package archive – Have CI produce a .tar.gz/.zip, push to Artifactory, fetch and unpack locally before dagger call.

How: Dagger loads from the unpacked folder, no Git at build time.
Pros: Zero Git calls during builds.
Cons: No live refs, adds a packaging step.

3. Caching proxy – Put Nginx or Varnish in front of GitLab’s repo archive endpoint.

How: First fetch of a commit hits GitLab; all later requests are instant cache hits.
Pros: Almost no repeat API calls.
Cons: First fetch still goes to GitLab.

Let me know which one you want to explore and I’ll be happy to assist