## Summary When a pipeline fails, it should open an issue and tag it with the new label (to be created) used to track Distribution pipeline failures. Currently, Distribution team members are responsible for creating an issue in the relevant project when a pipeline fails. For better group-wide visibility, we've also been tracking these in a Google Spreadsheet so we have some idea of common failures and how frequently they're seen. However, it's relatively easy for the team to miss any of these manual steps, and it takes time away from regular duties and especially triage duties, as the person on triage for the week is typically responsible for keeping an eye on pipeline failures. The team would find it very helpful for these issues to be automatically created, labeled, and filled in with as much helpful information as possible so they can be properly tracked. This way, we can create an Issue board and properly triage them. This aligns more closely with current processes and reduces opportunities for human error or missed data. ## issue-bot As a first iteration, we created [issue-bot](https://gitlab.com/gitlab-org/distribution/issue-bot). It's a small Golang program that will create an issue when a pipeline fails, logging the commit SHA, pipeline ID, and links to any jobs in that pipeline that failed. It uses scoped labels so we can create sortable Issue boards. It is bundled in a Docker image so it can be easily consumed in CI pipelines. ## Acceptance criteria - [x] All active Distribution projects contain CI configuration that will automatically create an issue when a pipeline fails **for the default branch on `gitlab.com`** - [ ] All active Distribution projects contain CI configuration that will automatically create an issue when a pipeline fails **for all scenarios that trigger a Slack notification** - [ ] Slack notifications are disabled, or modified to link to the issue created by issue-bot