We have an situation where the end-user get an exception that the query used more than 300 seconds.

After this the database session goes into wait event SQL*Net Message from client. Then after a while it ends up with wait event PGA Memory operation.
Why dosn’t the database session get terminated or an cancel query issued when the client call is terminated? Is it a bug in IFS 23r2 causing this behavior ?
Also can anyone explain the difference between this two time outs that are possible to set in the odata pods?