I am experiencing failures of JBPM upstream performance tests probably due to transaction timeout.
https://qe-jenkins-csb-business-automation.apps.ocp4.prod.psi.redhat.com/job/custom/job/mcervena/job/jbpm_upstream_problem/job/perf-jbpm/11/console

With mmacik@redhat.com we believe that the problem might be the sequential scan due to missing index on TaskEvent (processInstanceId):
https://github.com/kiegroup/jbpm/blob/master/jbpm-human-task/jbpm-human-task-audit/src/main/java/org/jbpm/services/task/audit/DefaultHumanTaskArchiveLoggerProvider.java#L43

As can be seen in the console output of #11 run it seems that after cca 70 seconds the transaction is aborted due to timeout.

The problem should be fixed by adding the missing index as can be seen (for postgresql) in the PR:
https://github.com/kiegroup/jbpm/pull/1973