Hi John,
I don’t monitor Platform Event itself. However every job is custom object, using reflection I deserialize them to submit job if limits are full (You can check my post on 100 queueable to see proof of concept). My whole error handling and reporting is around this object. I have dashboard which gives me insights of job and performance in last 30 days, 7 days and today. I have retry mechanism with some guards and rails. Example – I don’t do retry on validation errors. I do retry on Record locking error. I dont do retry on some classes, I de – prioritize few jobs on basis of class name etc.

It all started with basic code and eventually ended up with framework itself.