[Alerting] should log events when alert instances come and go

[pmuellr]

One of the design docs for the alerting history shows historic entries for when an alert instance starts scheduling action groups, where it wasn't before.

To deal with this, we'll want to have the alerting plugin write an event log entry when it notices an alert executor has scheduled actions for an alert instance that is not currently "active".

There has been a similar request for the opposite. Log an event when an alert instance has stopped scheduling action groups. In fact, we've talked about this being some kind of special action group (eg, "resolved") that we would automatically schedule when we noticed that happened.

Since I'm not sure we've seen the requirement for dealing with "new alert instances have appeared", figured I'd open an issue; please close if a DUP.

[elasticmachine]

Pinging @elastic/kibana-stack-services (Team:Stack Services)

[pmuellr]

This function was added in PR #61706, here:

https://github.com/pmuellr/kibana/blob/19fa8fa5121cbb961e593a4eeb6d521d6451ea40/x-pack/plugins/alerting/server/task_runner/task_runner.ts#L351-L358