So, here is a quick post but which wasn’t so eaasy to be mapped, though.
It happens that right after a 19c upgrade from 18c, we started facing Memory Leak messages related to PGA area on a regular basis. After a while following the v$process_memory allocation, it was possible to map it to an ODI routing with DBMS_STATS.GATHER_TABLE_STATS. The PGA exceeds the limit just a few minutes after.
Nice han, so what is happening?
Long story short: Bug 30846782 : 19C+ FAST/EXCESSIVE PGA GROWTH WHEN USING DBMS_STATS.GATHER_TABLE_STATS.
According to Oracle Docs: No Workaround!
But here is the golden info I want to share after an SR:
EXECUTE IMMEDIATE 'alter session set "_fix_control"='20424684:OFF';
You are welcome!