I currently have a Pega environment (v7.1.9) with multiple nodes. The standard configuration only has the SLA agent running on one node. I am looking to improve the processing performance of the system and it seems like one of the bottle necks we have is the processing of the SLAs.
The SLA agent is a standard agent, I am intending to enable it on all nodes. Is there any reason why this should not be done?
Thanks, that is good to know. I'm not sure we can justify have multiple dedicated only to running the SLA agent, this seems like an inefficient way of managing the problem. Are you suggesting that no other activity can be done on these nodes (i.e. remove them from the load balancer, do not allow front end access etc...). How would agents which run a different class (e.g. DefaultEntry) impact the SLA agent?
After loading 40,000 cases into the system overnight, we have noticed the number total number scheduled queue items have rocketed from 100,000 to over 300,000 items. This is almost completely split between SLA queue items and Incremental Index queue items. The problem is, because there are 300,000 items, the SMA does not respond when i try to open the scheduled list.
I can understand some of the SLA volumes, some of the cases have multiple SLAs (parellel flows) and will be treated as separate queue items, but i'm not quite sure why the incremental index queue is so large.
I appreciate I am going slightly off topic but any advice would be very helpful.