Determine an appropriate communication protocol for suspect incidents

During a recent issue: https://gitlab.com/gitlab-com/gl-infra/infrastructure/issues/6147

We were asked by a few members of GitLab about slowness, but no incident or specific person was paged to look into a suspect problem. When people suspect that there might be an issue we need a way for that message to reach the on-call person. Dropping a message in Slack and expecting a response may be against our values of asynchronous communication, however, it's still horribly important to get a message to someone if there's a problem that would negatively impact the larger community. Utilize this issue to track building a communication or protocol procedure to figure out how we can ensure that we meet the following criterion:

1. If anyone suspects a problem with our infrastructure, they have a way to reach us
2. Determine a threshold that way our Engineers are not overburdened with constantly checking 24/7 whether or not our infrastructure is suffering from a problem