[Stack Monitoring] Remove dependency on cluster_stats to load monitored products

[klacabane]

Summary

Today it is not possible for SM to display monitored products (eg logstash) without a corresponding cluster_stats document (generated by elasticsearch monitoring) that shares the same cluster id. this creates a strong dependency on elasticsearch being monitored for users only interested in a specific product monitoring and prevents the UI from loading when the cluster_stats document fails to be generated. Standalone cluster is an exception that works with hardcoded identifier used to trigger logic in both public and server codebases.

SM should be able load when elasticsearch monitoring is not enabled and other products are being monitored. The same way a standalone cluster appears in the cluster listing, any existing cluster_uuid stored in monitoring indices should generate a cluster object given at least one supported product is monitored.

Implementation details

Looks like we could replace this call that uses cluster_stats documents to build the cluster objects by a query that looks for any valid (ie generated by a supported product) document aggregated by cluster_uuid, we could potentially replace the standalone path with a missing aggregation.

AC:

• SM loads when elasticsearch monitoring is not enabled and at least one product is monitored

[elasticmachine]

Pinging @elastic/infra-monitoring-ui (Team:Infra Monitoring UI)

[miltonhultgren]

Is this the same as #125401?

[klacabane]

duplicate