Tuning of the Cassandra read time via the DDS Data Set
I am looking for a configuration to tune the read query time from the Cassandra table in Pega, while using the DDS Data Set to read the data from Cassandra. Is that possible by e.g. DSS configuration, or changing the yaml file? The client, that I am working for uses an on-premise installation with two Cassandra PODs (6 cores / 250GB each) configured, as the external DDSs (Decision Data Stores) in Pega. We are currently using a batch data flow run to read and aggregate ~150mln of the records from the Cassandra table and most of the data flow run time (~70%) Pega spends on a read operation from the DDS Data Set, so we would like to improve that read query time.
I would start by collecting more details of the offending queries, enable the debug log for "com.datastax.driver.core" - just create a log category with the logger name identical to the java package name.
I wonder if you have done any benchmark regarding your external C* cluster. Are you running this chart: https://hub.helm.sh/charts/incubator/cassandra. There is a stress tool there. That way, we can see how the C* cluster behaves even outside Pega.