Move to generic updates and handle gracefully fetching errors

1. Breaking change: preemptive_update binary sensor was removed since updates are coming now not only before alerts but also for de-escalation updates. There is a new attribute selected_areas_updates which should be used instead. A usage example can be found here:

2. More graceful handling of URL fetching errors: if there is a cached content of the URL it's being used. This can prevent the entity from getting into unavailability state for temporarily fetching issues.

Full Changelog: v2.23.0...v2.24.0