Our production environment was having single JBOSS cluster with 4 JBoss instances. As the user volume increased by 50% last week we added an additional jboss cluster having 4 jboss instances and used SSL PORTS (443 & 444) to manually route user requests to one of these clusters. If the PORT is 443 F5 will route the request to first cluster other wise to second cluster. But after adding the additional cluster production system started behaving odd, performance got impacted and the second cluster went down after couple of hours. Our JBOSS admin identified that as the work index is on one of the instances on first cluster most the time the requests from second cluster also being redirected to the instance having work index.
Does pega support work index on instance on each clusters to avoid this issue? Or what is the best solution to address this issue