As an Evergreen administrator, I want to be able to ignore data coming from stale instances,
so that the logs and data coming from these instances don't end up hiding the current issues on active and up to date instances.
I also want to inform the users of this instance that it cannot be updated automatically anymore correctly, hence they are possibly vulnerable to attacks, and provide them a path to reach out to the Jenkins Evergreen team.
Even when we fix issues in Evergreen, Jenkins core, a Jenkins plugin, we keep seeing many issues getting automatically reopened and marked as regression by Sentry.
For example, we see an instance that is still stuck at Update Level 66 (current last is 174).
So I think we need to attack this from different front:
- Core goal: add a way to mark an UUID as ignored, so the Evergreen backend can then refer to this
- immediate goal: not forward their logs anymore to Sentry,
- maybe later: reject connection possibly in the future, etc.
- From the instance, detect that is outdated, and surface a warning to the end user that the instance they are using is outdated. (Likely through an AdministrativeMonitor in the Evergreen Jenkins plugin)
- Link to a page somewhere (likely a redirect on jenkins.io) explaining how to reach out and provide data to understand the cause of their issue, and see how to resolve it