I have recently implemented a fix/workaround for this a few weeks ago which has resolved every single MBEP issue I've ever had since purchasing the product last September. This includes endpoints randomly dropping off the cloud console and also causing interference with our regular AV and other unrelated house software applications (which were white-listed). None of the suggestions from the support staff has ever worked for us, including those listed above. Since implementing my own workaround, our MBEP struggles have completely stopped.
The issue apparently is caused by a persistent memory leak in the MBAMService.exe process. Upon endpoint startup, the memory used (in Task Manager) on all our PCs starts at around 250,000 K. When left unattended, that memory usage will slowly creep higher and higher. After a few days to a week, that memory rises on our endpoints to 400,000 K to 500,000 K at which point we start having issues with our other software. Also, at this point, we start to see our first set of endpoints disappear from the cloud console. More and more endpoints begin to disappear as the hours & days pass. If continued to be left unattended, the memory rises to between 500,000 K and 1,000,000 K. At this level, MBAMService.exe CPU usage rises and gets stuck well above zero. On our fastest machines, the MBAMService.exe CPU will run steady at 13%, just enough for our users to notice performance hesitations. On our slowest machines, it will run steady at 50% in which the machines become basically crippled. To stop this, the endpoint must either A) be restarted, or B) the MBAMService.exe process must be manually killed and restarted. After doing this, everything calms down and starts working again. Everything.
So after months of endless frustration, I ended up writing a .cmd script which stops the cloud service (MBCloudEA.exe), forcibly kills the MBAMService.exe process, restarts MBAMService.exe and finally restarts the MBCloudEA.exe service. I put a copy of this .cmd script on the endpoints and set Task Scheduler to execute it (as Local System) at 6:00AM every morning. Ever since implementing this workaround, all our MBEP problems have completely vanished and I haven't looked back since.
If interested, here's the script I'm using:
:: Reset Malwarebytes Endpoint Protection Services
NET STOP MBEndpointAgent
TASKKILL /IM MBAMService.exe /F
TIMEOUT /t 10 /nobreak
NET START MBAMService
NET START MBEndpointAgent
It's incredibly simple. I chose to "kill" the MBAMService.exe process rather than stop the service because if the high CPU usage gets stuck, it has a problem simply stopping the service. Upon killing the process, MBAMService.exe re-executes itself, but I manually start it again in case it doesn't. Restarting with it already running doesn't harm anything. Also, if the issue is allowed to get bad, the Cloud Service has problems and sometimes stops on its own when MBAMSercvice.exe restarts. That's why I stop and restart the Cloud Service at the beginning and end of the script.