Just wanted to log in and share my experience as well. It's 100% equal to what Matthew had indicated. I also am running an on-prem NG virtual appliance on version 19.5.1.7.
The problem started halfway through December where I would log into hivemanager and it would say "Unable to access data". When I reboot, it eventually started showing me data and I could navigate HM. That said, within an hour it'd break again, showing "Unable to access data".
At this point I opened a case with Aerohive and the tech had indicated it could be an issue with the fact that we had the VM on SAN storage. We had a call scheduled for him to do some troubleshooting by adjusting things like heap size.
Then I found Matthew's reply, so I spent a few hours yesterday browsing through the log files and deleting them. I tried to do a quick "Purge All" but it gave me a red pop-up error that filled the screen. I then made the page show me "500 at a time" and started off by deleting anyone that had a red status by using the "Delete selection" option. Eventually, just as Matthew said, despite the fact the resource monitor indicates our resources were fine, it was behaving as if resources were in fact the issue. I continued deleting files 10 at a time for an hour, but eventually I could check off the entire 500 log files and delete them as a selection. When done, I restarted the services.
Since then it's been 16ish hours and I'm still able to log in and see the polled data.
