Daily intermittent performance issues in Model-Driven and Portal Power Apps

Hi! I've been working with support since April 17th, 2024 to determine the cause of intermittent, but daily performance problems on the load of pages, case updates, and case submission in both Model-Driven and Portal applications in production. As of yet, support has not been able to provide a possible explanation or course of action for this problem.

Internally, we've ruled out, we believe, any issues with our internal network; both our internal network and connection to our ISP is more than sufficient for our traffic. There is no consistency between type of case form, the time of day, or user.

Example behavior:

Occasionally, at any point of the day, a load of any case may suddenly hang for over 30 seconds. In this network trace, we see several GetClientMetadata calls...

...that wait a long time (>11s) for server response...

Any ideas as to where I could start in troubleshooting this problem? Note that our Performance Insights is not working at all and MS has declined to look further into the problem as Performance Insights is in preview.

Has anyone used Application Insights for debugging such problems? Is it appropriate for debugging such problems?

Any help would be greatly appreciated. Thanks!

P.S. I've attached the har file generated from a network trace in Chrome if anyone is feeling ambitious.