Handle indexer cluster down in web interface #2713
As described in #2623, when ES is not reachable, going to the indices page in the web interface only renders a grey page with a loading spinner.
This PR handles the situation when ES is not available and displays a warning dialogue with some information.
I also saw that the ES cluster health was not being refreshed in the system overview page, and now we update the information when the
When mounting `IndexerClusterHealth`, refresh the information available in `IndexerClusterStore`. In this way we can display an accurate ES status when rendering the system overview page, without constantly asking for the health information.