I tried to fix this by trying to update each remote cluster in sequence directly in the UI, but it becomes quite messy given that we have a refresh rate of 1minute to refetch all the remote clusters again (in order to keep them up to date in the UI).
With https://github.com/elastic/kibana/pull/183593 we enhanced the connection status of the remote clusters list by calling the resolveCluster API. This can sometimes become problematic as the API has a very long timeout when reaching out unresponsive hosts. I've created https://github.com/elastic/elasticsearch/issues/114020 to see if we can have ES give us a way to configure the timeout.
I tried to fix this by trying to update each remote cluster in sequence directly in the UI, but it becomes quite messy given that we have a refresh rate of 1minute to refetch all the remote clusters again (in order to keep them up to date in the UI).