Something that we are looking at more an more is the ability to have a controlling system that is able to perform disable, terminate or whatever actions on an unhealthy host. I.e. AWS instances that seem to start to degrade fairly regularly.
This is something that is currently available where values are compared against trigger expressions but this is currently done at the server. What we are proposing is that actions can be taken much closer to the host to as to reduce the dependencies on the connectivity back to the server.
Easiest place for this to happen would be the proxy but the most resilient place would be at the agent level.
This seems to my like it should be a module because it would need to be probing the buffered data at the particular layer.
The trigger table records for all the hosts that are monitored by Proxy would need to be included in the proxy config sync
The agent module would need to include trigger details when it asks the proxy for it's active checks.