Unified Event Notifications

We clearly have an issue with the myriad of ways notification
events are being handled in Nova.

We have:
1. The AMQP notifications,
2. The DB InstanceFault table,
3. Exception handling decorators,
4. The conductor mechanisms you've highlighted,
5. and, to a lesser extent, logging

There is a mismatch of data stored in the InstanceFault table and what is sent to the notifications. (why do it twice?)
Then there's the problem of squashing a private data leakage problem in
one place to have it appear in some other place.

Ceilometer is working on shoring up #1 to provide a
meaningful data package, but there is clearly duplicated effort.

Is storing ever-growing InstanceFault data in the
production database a bad idea? Should our rule should be: "Get archival
data away from production Nova as quickly as possible."