The issue is that on each card addition, an api call is being made to grab the valid previous activity outputs. Originally, this call contained way too much data (choice list data).
This causes each card addition request to freeze the browser render thread, causing unusable features and can lead to the instance being locked up as network requests wait on the previous one to come back.
This extra data was removed from the jelly dot walkers in Quebec but the change did not apply to the client side API call at the time. This was fixed in Rome so a back port is necessary into a patch in order to get these fixes ported over.
Steps to Reproduce:
1. Login to a quebec instance and navigate to Process Automation -> Process Automation Designer.
2. Create a new process.
3. Add lanes and start to add create task activities in rapid succession.
4. Try to add more activities to the lane and you would experience the delay for every single activity you have added.
5. After a while, the following error will be displayed:
"Error: A network error occurred. Please refresh the page or contact your System Administrator if this problem continues."
6. Try to refresh the page and notice that the instance becomes irresponsive. The screen becomes just blank without loading anything and will be back to normal after a while.
There is no reasonable workaround at this time.