Testing below will help clarify the issue seen of missing metrics within the dashboard and references below also provide steps for users who may want to test this behavior.
This article aims to address the confusion users experience when seeing the graphics with "missing" data points and clarify that this is expected behavior.
For more information on how to further narrow the scope of your issue, refer to Troubleshooting App Metrics. If you are still experiencing concern or questions, engage with Tanzu Support.
To replicate the behavior and/or provide a reference to the expected results in Metrics, deploy a test application with multiple instances, in this case it is 4 AI's.
Test
For example, continuously send requests to the app without specifying a particular AI:
date ; while true ; do curl mytestapp.cfapps-06.slot-59.###.#####.com/id -H 'Cache-Control: no-cache'; sleep 5 ; done Thu Aug 26 16:18:09 EDT 2021
Observing through Metrics shows metric data populate within our App Metrics graphic. As this is aggregated by process (Web), users see the graphic show consistent traffic until there is none destined for the app (or we stop curl requests).
The aggregated output shows that there is missing data between when the requests stop and start again.
When stopping curl against all AIs instead of targeting the single AI (instance 2), the graphic displays expected behavior with "missing" requests. Application traffic is reported as non-existent on any AI except AI 2.
For example, you'd use this command to stop curl against all AIs instead of targeting the single AI (instance 2):
date ; while true ; do curl mytestapp.cfapps-06.slot-59.###.#####.com -H "X-Cf-App-Instance":"8edcde2a-####-####-9f8c-c1999d13aec9:2" -H 'Cache-Control: no-cache' ; sleep 4; done Thu Aug 26 16:33:11 EDT 2021