I have been consistently facing the same below issue since I started working with site workflows (kicked-off manually).
I have a site workflow that queries 2 lists and then is supposed to create about 5k items in a 3rd list based on info taken from said 2 lists. Following the best practices/debugging practices I have seen in this forum and elsewhere, I implemented pauses:
- 1 min at the very start of the workflow, before it even does anything
- 1 min after each 200th iteration/item created
What I have noticed:
- the first pause sometimes (50% of time) taking as long as 20 min instead of 1 min
- the first pause at the 200th iteration almost consistently (80% of time) taking about 10-15 min and the workflow consistently failing after that pause is complete with nothing more than "An error occurred in..."
We have a lot of other list workflows for which I have never seen this issue.
Any help with troubleshooting the issue would be much appreciated. I have seen somewhere that it is something to do with the queue timer and that disabling/re-enabling Nintex at the site level might help. This is probably not be a viable option for us considering the farm is also hosting 48 other prod site collections.