https://www.atlassian.com/software/jira

Hopefully this would be easy to implement as deployment might be quite similar to Atlassian Confluence?

@girish The request, or maybe more of a pious wish, is for the KAS to be available within the Cloudron app itself.

My understanding regarding running a separate KAS is that this would effectively require spinning up a GitLab Omnibus instance, thus sort of defeating the purpose / convenience of the Cloudron instance. There doesn't seem to be a standalone KAS binary.

The request for the container registry (https://forum.cloudron.io/topic/1837/gitlab-how-to-run-container-registry/25) is somewhat similar.

GitLab supplies various DevOps tools to manage and run CI pipelines and even production deployments.

This is most convenient when connecting GitLab to a Kubernetes cluster. GitLab can then self-provision pods as needed to launch CI runners, deploy applications, run red/green deployments, and so on.

Starting with GitLab 14.5, the preferred way of managing such Kubernetes clusters is with a GitOps approach using GitLab's Kuberentes agent called KAS.

This is a separate service which the Kubernetes cluster can poll to pull in information and infrastructure specifications from GitLab.

Other methods of managing a Kubernetes cluster have been deprecated in GitLab 14.5. and will thus be removed in future versions (https://gitlab.com/groups/gitlab-org/configure/-/epics/8).

The documentation for enabling KAS can be found at https://docs.gitlab.com/ee/administration/clusters/kas.html.

However, setting

production:
  gitlab_kas:
    enabled: true

in /app/data/gitlab.yml seems to have no effect in Cloudron.

It would be great if the Cloudron package could be modified to allow for a KAS to be started.

Alternatively, there's also https://github.com/localtunnel/server (docker image available).

Already supports docker:

https://docs.tuleap.org/installation-guide/docker-image.html
https://hub.docker.com/r/enalean/tuleap-aio/

https://www.atlassian.com/software/jira

Hopefully this would be easy to implement as deployment might be quite similar to Atlassian Confluence?