[Response Ops] Refactor alerting task runner - refactor determination of active/new/recovered alerts

[ymao1]

Meta issue: #124206

As a step towards refactoring the task runner, move the determination of new/active/recovered alerts to the AlertFactory. The AlertFactory contains all the information needed to make this determination. We should be able to add functions to getNewAlerts(), getActiveAlerts() and getRecoveredAlerts() and store it in the alert factory so even if they are called multiple times, the calculations are only done once. Keeping in mind that we should not be exposing these functions to the rule type executors (we want them only to use create() and done()). We should also see if we can move the functionality in trackAlertDurations into the alert factory so getActiveAlerts() returns alerts with updated duration and getRecoveredAlerts() returns alerts with updated duration and end time set.

[elasticmachine]

Pinging @elastic/response-ops (Team:ResponseOps)

[ymao1]

Unblocking as #131544 is close to getting merged