Make sense of safety events

[joshmackwilliams]

The safety manager should take the incoming safety events and combine them to identify the root cause. For example, if one node says that all of its dependencies are down, the problem is likely with that node. On the other hand, if all of a node's dependencies say that node is down, the problem is certainly with that node.

Essentially, this issue involves taking all of the data flowing into the safety manager and turning it into something useful.