I created a job that runs daily (every midnight) that uses DBCC CHECKDB on every database. Every day in the morning I launch SSMS to look at the logs and hope to see CHECKDB found 0 allocation errors lines.
But if I pass some days without checking the logs and discover that some database is corrupted, how can I know when the corruption ocurred to restore the backups from that point in time?
Check in the table msdb.dbo.suspect_pages. It has useful information about which pages are suspect, although the date in there shows you when it was last seen to be suspect, which isn't exactly what you want.
Have a look at what type of page it is that's corrupt. If it's a page for a nonclustered index page, then just rebuild the index. If it's a table, then do a page level restore. You will know straight away if you restore that page from a backup where it's also corrupt, because the restore will fail.
Your best bet is really to look at your history of DBCC results.
