We are on Kinetic 2023.2.11, but primarily use Modern screens. A couple of weeks ago, we were noticing some terrible performance within Epicor for a majority of our users. We opened up a ticket and Epicor did something on the back end which resulted in kicking all of us out and preventing us from getting back in. After a few more minutes, we were able to log back in and our performance seemed to be back to normal, however the next week, we experienced the same issue and went through the same process of them doing something on their end, us getting kicked out, and having to wait in order to log back in.
After this instance, the person assigned to our case said that Epicor “identified the root cause which is backend noisy neighbor” and they are currently working on something to prevent it from happening in the future.
For about a week, half of which the company was off for July 4th and the weekend, Epicor seemed fine. However, these slowdowns started to occur again, but only for select users. We were able to relieve their issues by deleting the C:\Users[user]\AppData/Local/Epicor folder. However, for one user this has had to be done each of the following days and just now, I had to do the same thing for another user. Only about 4 users have reported these slowness issues while the rest of the company seems to be doing fine (or maybe they are just dealing with it???).
Has anyone else experienced Epicor performance issues isolated to just a few users?
EDIT: We are on the Cloud using Desktop Clients… just to clarify
I can’t speak specifically on why you guys are having the issues, because we were wondering the same a few months ago when we first started to transition and training up our user base on Kinetic. After a few weeks we started to notice a lot of performance issues for specific users as well with no real causation.
We submitted a lot of different tickets and spent a fair amount of time trying to resolve the issue, also by deleting the local cache for Epicor folder but continued to have issues intermittently every few weeks.
What we ended up doing was switching all of our users to the browser instead, as we are Cloud also. We were having a lot of issues with the desktop application and since making the switch all of the performance issues we had have since been much better and mostly resolved.
I don’t know if this is an option for you guys, but it worked a lot better for us than spending the time chasing the issue and waiting to hear back from support and the elusive “backend changes”.
Unfortunately, we have not had much success with Kinetic screens so don’t think we can switch to the browser just yet. Hopefully we can move in that direction in the future as I have heard from others that browser mode is pretty smooth.
My bad on that, I forgot you had mentioned that you guys are using the Modern screens rather than Kinetic. The browser has been pretty good to us so far, I can’t complain (as much)
I’ll keep an eye out for anything that might help you guys, and hoping to hear that it improves in the near future if support can get it figured out.
OK - I think Client Tracing (with include Server Trace) might be one place to start looking.
Do the Server times get high, or are the times since last call high (when analysed with the PDT).
With the Times since last call, these can be high because the users is sitting doing nothing in which case ignore. OR - in cases where the client is stringing together a whole list of calls, and these get high - then the client is misbehaving - could be due to weird caching, anti-virus exclusions not working or using Windows SSO (which is how I discovered that issue).
Should work for cloud instances too. Might need to go to epicweb to download it. But it will take a client trace and make it into a prettier table - so no reason it shouldn’t work with cloud.
