[BUG] Grouping issues in the Overview screen

[cakrit]

We have a few problems with the groupings shown above in the overview screen:

• "group" should be renamed to "cgroups"

• "cgroups" are missing the cgroup network interfaces. e.g. see the following container, it has a "net eth0"
  

• "prometheus" is NOT a group. For each endpoint that we collect data from (e.g. redis), we need a separate grouping.

[cakrit]

@jacekkolasa please speak with Ilya about this, it should be easy to fix.

[jacekkolasa]

I have limited options on how to group prometheus collectors, because on composition step we loose the app information, so the only way I can group is by job.

This is how charts are named in node-view:

prometheus_prometheus-mongodb_mongodb-sharded-shard0-data-1_metrics_tcp_9216.promhttp_metric_handler_requests_total
prometheus_prometheus-mongodb_mongodb-sharded-shard0-data-1_metrics_tcp_9216.promhttp_metric_handler_requests_in_flight

And in overview:

prometheus.promhttp_metric_handler_requests_in_flight
prometheus.promhttp_metric_handler_requests_total

@cakrit is this intentional? I guess it is, but I want to confirm.

This is what we get when grouping by job, which i get from first part of the family, for example:
promhttp_metric > promhttp

Is this OK?

[cakrit]

This grouping is good! I'm closing and expect to see it live soon.