We have exposed Appian WebAPI to external and when it's being invoked like 1000 times in a bulk load process, and the linked Process Model itself just writes to 4 Record Types in Appian. But what is being observed is that, the Process instances are getting stuck at various nodes (Sometimes in XOR Node/ Some times on query Record) and on an average we are seeing 50 instances getting stuck out of 1000 being invoked. But if we only have 100 API calls at a time, then this isn't an issue. Any experience or suggestions.
Discussion posts and replies are publicly visible
Starting a higher number of processes in a short period of time induces severe load on the system. I suggest to talk to Appian about appropriate system sizing.