Even in cases with hundreds of executions, the user or sys time being zero is not correct. Therefore, this needs more review. This could be an issue any because of poor handling of query nesting.
It seems, due to same problem, for INSERT queries at times PGSM shows (cpu_sys_time + cpu_user_time > total_exec_time)
Even in cases with hundreds of executions, the user or sys time being zero is not correct. Therefore, this needs more review. This could be an issue any because of poor handling of query nesting.
It seems, due to same problem, for INSERT queries at times PGSM shows (cpu_sys_time + cpu_user_time > total_exec_time)
https://jira.percona.com/browse/PG-612