Users in your PROD database complain about the slow response of the transactions. While investigating the reason, you find that the transactions are waiting for the undo segments to be available…

Raising undo retention instructs Oracle to preserve undo data longer, which would consume more space in an already constrained tablespace and worsen contention, not resolve it.

Under Automatic Undo Management, Oracle manages undo segment creation automatically; manually adding segments is unsupported and does not address the space shortage.

Only one undo tablespace can be active per instance at a time, so creating a second one does not add usable space to the active tablespace experiencing contention.

With undo retention at zero, Oracle expires undo as soon as it is no longer needed by the generating transaction, meaning waits for undo availability point directly to insufficient physical space in the tablespace rather than a retention policy problem. Increasing the size of the undo tablespace provides additional extents for undo segments to allocate, removing the space-driven contention and improving transaction response times.