We have a system where we define alerts on mount point getting full rather than tablespaces getting full as these alerts dont include aut extend on and are sometime deceptive. In our system we have datafiles which can grow max to 25g including autoextend on , can some provide me a query / script or alert to be set in oem which could give accurate alerts or mails on tablespaces getting full at around 95% including the autoextend on.
Assuming you are referring to the Tablespace Used (%) metric, this metric takes into account auto-extensible data files and available disk space. Note that for pre-11g databases, you need database patch 5549540 to exclude objects in recycle bins from space used calculation. If necessary, look at MOS 403264.1 on how to troubleshoot tablespace used alert issues.
Hope this helps.