Resolve how to handle case when the handler crash

it can be solved in many ways.

An option would be to have somehow kind of redundancy on the daemon itself which will respawn and monitor a sub-daemon.
In case it fails it will be respawned.

Since the alert-handler can be fondamental component, I do think that somehow a similar solution where only this binary is involved for its own redundancy/resilience is the best solution