This is an epic parent task. Actual work will be done in subtasks.
Java repos
- Wikimedia Event Utilities - https://gerrit.wikimedia.org/g/wikimedia-event-utilities
- Analytics Refinery Source - https://gerrit.wikimedia.org/g/analytics/refinery/source
- Hdfs-Tools - https://github.com/wikimedia/hdfs-tools & analytics/hdfs-tools/deploy
- gobblin-wmf - https://gerrit.wikimedia.org/r/plugins/gitiles/analytics/gobblin-wmf/
Deployment of these repo's jar artifacts to GitLab is tracked in T367405: Migrate existing Java packages to deploying to Gitlab, including new version of parent pom, validation that all dependencies are available, and validation that deployment to production still works. Ideally we could migrate these to GitLab before work is done in that ticket to make Jenkins deploy to GitLab. That way the work to deploy via CI only has to be done once in Gitlab CI.
Other repos
Note: scap & 'deploy' repos are listed here, but we may not need to migrate them if we change our deployment process. See T322690: Add support for repository artifacts in Airflow
- analytics/refinery & analytics/refinery/scap (T362832 is related)
- -- This should include making refinery-python a proper python package and using GitLab CI to publish wheels and conda envs as needed. T382703: refinery-python should be moved to the data-engineering namespace
- wikistats 2.0 (Test Kitchen perhaps?)
- analytics/turnilo/deploy - Q: do we have plans to host turnilo in k8s? If so we may not need to migrate this.
Other work
- Once wikimedia-event-utilities repo is in GitLab, we would like to merge eventutilities-python into it, and release the java and python artifacts together.