Hi,

I tend to set the trace log to 6 when the system is slow as a first step and then search the server log file for ‘Time Elapsed’. Locate the transactions that are taking in the seconds or just vastly longer than others and see if they relate to a common area.

As general advise, the activity feed tends to take up a lot of process and many users/customers do not actually use it. Turning this off can give a big performance boost with little noted change.

Hope that helps a little.

Lee

Thanks Lee for answering.

We can not enabled the trace log in Production Environment because we are afraid that it will degrade the performance. However we have already disabled Activity Feed from the beginning. Moreover we have some monitoring setup in Appdynamics where frequently we can see the below error across three of our different servers at different time slot.

“CLR Garbage Collection Time is too high”

Could you please help on this?

The trace log should have very little impact on server performance. When I have used it in the past it has made an unnoticeable difference so you should be okay to turn it on, especially if only for a couple of minutes to capture data and then switch back.

A more detailed analysis would probably be better suited to a support case so, if you are reluctant to turn on the trace or the trace does not reveal anything, I would advise creating a new case with as much information on when and how you are encountering slowness in the system.

Lee

Just confirming what Lee said, Trace log (log level TR=6) is ok to use in Production.  We’ve had customers using this level in Production for weeks and months at a time.  Don’t increase any of the other log levels, just TR=6.

I recommend turning it on for just an hour at first and collecting the server logs during that hour long period.  The logs will show the elapsed time for each XML message sent to the server.  You can then identify which messages are taking a long time.  Turn It back down (TR=0) after collecting the logs.