APM's metrics tab doesn't show when using Kibana spaces

Kibana version: 7.4
APM Server version: 7.4

We have a different elastic cluster for each environment – staging, production, etc. Only one Kibana instance is used to visualize data from these clusters using Spaces for the different environments. But the APM dashboard doesn't work out of the box and expect data to come from the apm-* index patterns. What we have configured is to search for *apm-* which worked and transactions are showing, but for some reasons, metrics(cpu, mem) don't show up. We have resorted to using another Kibana instance only for the APM at the moment.

Is there a way we can configure APM dashboard to catch metrics when using the *apm-* pattern.

Note: I'm not the person doing the server configuration, so it might take me some time to figure out all the details and test.

Hello @Evgeni_Dzhelyov! Couple of questions:

  • Just to be sure, when you say APM dashboard, do you mean the APM UI/application inside of Kibana? In previous versions, we also offered APM-specific dashboard configurations for Kibana. See https://www.elastic.co/guide/en/apm/server/current/load-kibana-dashboards.html.
  • Are you not seeing the metrics tab, or is the content of the metrics tab not displayed? Do you see a blank page?
  • What kind of agent(s) are you inspecting? The RUM agent will not show the metrics tab initially, and from Kibana 7.5, the same will apply to the Java agent (which will display a JVMs tab).
  • How did you configure the APM UI to read from other indices? If via kibana.yml, you'll need to configure apm_oss.metricsIndices as well. I just noticed that we are not mentioning those in the documentation, we'll make sure to get that fixed. https://www.elastic.co/guide/en/kibana/current/apm-settings-kb.html

Hello @dgieselaar, thanks for such fast answer!

  1. When I refer to APM dashboard I'm referring to the APM UI application inside Kibana. What is the appropriate name to refer to that from now on - APM UI?
  2. The metrics tabs is present, there are metrics in the index, but the tab is not showing any data. If we switch to another Kibana instance, without Spaces and configured for that cluster, there is data.
  3. I'm inspecting the Ruby 3.x agent and as I point in the previous point (2) there is data.
  4. I'm pretty sure it's configured in the kibana.yml on the server to read from *apm-*. I will double check the configuration of apm_oss.metricsIndices and write back to you.

As a side question, I see the ruby agent report various metrics including GC and heap memory, but the metrics tab display only a general CPU and process MEM. Is there a way to see the more detailed messages from the ruby agent in the UI?

What is the appropriate name to refer to that from now on - APM UI?

Yes, we use APM UI. I can understand the confusion though :slight_smile:

As a side question, I see the ruby agent report various metrics including GC and heap memory, but the metrics tab display only a general CPU and process MEM. Is there a way to see the more detailed messages from the ruby agent in the UI?

We have some Kibana dashboards for the Ruby agent, and they should include charts for both GC and heap memory metrics: https://github.com/elastic/apm-contrib/tree/master/apm-agent-ruby/dashboards