Auto Log Alerts
Most logging frameworks allow definition of logging level, for example
fatal down to
trace. With the foundation of gitlab-ce#37769 (closed), we can begin to offer advanced warnings about changing log behaviors which may not be surfacing elsewhere. (Metrics, etc)
We could build some automatic parsing of these log events, and for serious logging statements (perhaps
error or above) we could begin to track the rate of generation.
If there is an abnormal number of these events than usual, alerts can be generated.