[Monitoring] More resilient against errors with alerting

[chrisronline]

Currently, if there is any kind of error with fetching alerts, the Stack Monitoring UI fails to load (as we don't explicitly catch any errors when fetching alerts when loading all clusters) which shouldn't happen.

[elasticmachine]

Pinging @elastic/stack-monitoring (Team:Monitoring)

[chrisronline]

Reopening this as we need to make the /status endpoint more resilient as well.

[crespocarlos]

closing in favor of #189154

[crespocarlos]

closing in favor of #189154