Related to elastic/kibana#104986
Follow up after #3051 (comment)

In the original plans of remote ES output, we planned to make Fleet Server status degraded if a remote ES output is not accessible, however this has too large impact, so we decided to do output health reporting separately.

Implementation tasks:

• Report periodic health updates per remote output to logs-fleet_server.output_health-default data stream or metrics-*
  • Check if privileges are needed in ES for fleet-server service account to write to these data streams, and kibana_system to read it (metrics privilege already exists for kibana_system)
  • Add ILM/DLM
• From Fleet UI, read this data stream and show error status on Output UI