[APM] Bug: No data empty screen persists even if there's data available

[formgeist]

Kibana version:

8.0.0

Original install method (e.g. download page, yum, from source, etc.):

APM Integration testing using the following compose;

scripts/compose.py start 8.0 --all --with-elastic-agent --apm-server-managed

Description of the problem including expected versus actual behavior:

The no data empty screen appears for APM and User Experience apps although there are service data available within the last 15 minutes. I expect the service inventory to show and the user experience app to be able to show RUM data.

Note: I initially thought that the index pattern might not have been loaded properly because there was no APM related index patterns to choose from in Discover, so I recreated them by "loading Kibana objects" in the add data tutorial. Didn't fix the issue though, but I was able to look up data in Discovery.

Steps to reproduce:

1. Go to APM or User Experience
2. The default no data empty state shows instead of the service inventory view

Additional details
https://elastic.slack.com/archives/C877NKGBY/p1639442846469800

[elasticmachine]

Pinging @elastic/apm-ui (Team:apm)

[muneebsajjad]

@formgeist for an early contributor what should be the starting point in debugging this issue!, moreover do we have any high level architecture diagram of the system to understand how multiple modules/components are integrated together... which might help to understand the codebase as well.

[formgeist]

@muneebsajjad Thanks for your interest, but I think this issue to better left to the @elastic/apm-ui to resolve as it's a bug in an upcoming Kibana release. We don't currently have much documentation on the app architecture other than what's presented here in our Kibana docs and in the plugin documentation itself. I would recommend looking out for another less urgent issue to help fix.

Thank you for your contributions already 🙌

[formgeist]

I'm trying to reproduce this bug using the script for the apm-integration-testing above but no luck (containers are coming up unhealthy). Will try and investigate further this week and see if I'm getting anywhere - if I cannot reproduce, I'll close it out.

[formgeist]

I've not been able to reproduce this error using Cloud staging and APM integration testing combination to run agents locally to a managed apm server on Cloud staging running 8.0.0-SNAPSHOT. Closing