Trigger monitor frontend sucks

[coderdj]

The trigger window needs a bit of an overhaul. There's good info in the backend but the decision to dumb it down limits ability of shifters to diagnose issues. A few ideas:

• The RAM in each machine is 128GB (of which maybe half is available as a mongo buffer?) while the SSD in each machine has ~880GB usable as the mongo buffer. The buffer monitor should reflect this, also with a help dialog that explains in which case each should be used.
• Some more useful tracking of the deadtime would be helpful
• The overall health thresholds should probably be maintained, but with better descriptors for what's wrong and more reasonable thresholds in general

[coderdj]

0d9fd26 makes a bit cleaner. Backend needs a bit of love still