Document a process for communicating CI test failures on `main`

[aaronsteers]

Very infrequently, we'll see a merge to main that causes one or more CI tests to start failing - either consistently or sporadically.

The team generally gets to know about this through osmosis and engineering threads in slack, but it might be missed by other community members or team members.

This proposed process change would be something like:

Any time a team member becomes aware of a CI test failure, I think we should perform the following steps:

1. Immediately log an issue to log the instance of the CI failure, including mention of it is constant or sporadic.
2. Share the link to the issue with an FYI to the #contributing channel.
3. If the PR that introduced the error is known, assign to the author of that PR to request a resolution weight estimate. Otherwise ping a manager for next steps.