Normally I would have "abort tran on log full" option set on all databases.
This way it reduces some aggravation.
Also needed is monitoring of open and long running transaction visible from master..syslogshold.
This should tell you who is most probably causing the issue.
Most of the shops use combination of scheduled and threshold triggered transaction backups.
This way any abnormal high activity can be also be addressed.
10 minutes could be too long for a hungry update affecting millions of rows.
I have seen anywhere from 10% to 25% for log space.
But it will depend on your local situation/application profile etc.
Do NOT fall in the trap of increasing log space whenever users ask. Instead ask them to break down the transaction in manageable batches/chunks and commit often. Log space -- just like tempdb --- is a shared resource and each connection needs to play by some acceptable rules !!
HTH
Avinash