We have an implementation that is partner hosted. Currently we have about 75,593 system jobs that are either reporting error or the dreaded waiting for retry due to error. These errors go back 3 years so we are averaging about 500 errors a week. My hosting partner says these errors are normal but I see plugin exceptions all over the place. I am constantly running workflows to manually to push things through the pipe because the original workflow failed, so I assume its not as normal as they would like to think?
Some of them are legit, like missing email address but filtering those out I still have 71,217 errors of one kind or another. I do see some simple workflows that are failing that have very little logic. Things like clearing or updating a field. I see some errors that are reporting records not found but when I do a record search by GUID (XRMToolbox), the record is found and the created date is well before the reported error or when the process was triggered.
Now I can dig into these (Looks to be 209 workflows) but I am thinking a CRM trace would be a better place to start to try and get an overall health check of the system. Am I wrong in assuming that? I am thinking the asyncoperationbaseshould should be dumped and a trace enabled? I have done that before with success. I have never really seen an instance with so many errors but I could be wrong and maybe this is the new normal for Dynamics CRM.
We are about 20 users with maybe 10 of those being heavy users. We have maybe 20 members logged into the portal at any given time.
The final thing on this is these workflows do not fail all the time. Its just some of the time, so its not really the workflows that are the issue I think. I think its performance tuning and maintenance or lack there of?
Thoughts?
