As a follow up to https://issues.redhat.com/browse/LOG-953 we want to configure it such that we have some 10000 ft view of index level metrics (e.g. number of namespaces and maybe number of records for each namespace)
So that a customer can see which namespaces may be the most noisy.
- We are able to gather and view index level heuristics within Prometheus
- Ensure that we can see metrics split out per namespace
- Ensure that we are displaying current document size and log rate for each namespace
- FYI: We haven't productized the mapper-size ES plugin  that is required to get size of individual documents yet. Hence this goal hasn't been reached yet.
- We will not be configuring alerts as part of this
- Making sure that the plugin is generic enough that it can accept queries to execute
- This will be useful to have, but will not be implemented as part of this card – follow up to be created
Notes from spike: https://github.com/openshift/elasticsearch-operator/pull/626
Potentially additional metrics:
- Number of indices
- Number of shards