Need help improve the Group Aggregation operation

In IIQ8.4p1 environment, the Customer deployed webservice connector to pull user accounts and roles data from their DBaaS Oracle application. For the partitioned account aggregation (PAA), it takes less than 2 hours for the aggregation (145K user accounts scanned from 933 databases), however for the group aggregation, it takes more than 12 hours (scanned 398K Oracle roles from 933 databases). The customer wants to improve the GA operation and asks the same as the PAA operation. Please help me improve the GA with better performance. Any pointers would much appreciated.

Hi @jxu11,

you can improve the resource of server and JVM so as to be able to increase the threads for each machine or add a task server.

Also, you can use Delta aggrergation if the connector allows it and set the Partitioning on group aggregation task.

Otherwise you can unaggrergate the inactive group or the not necessary group, but depends of customer managment.