Originally posted by reydp
It's way too small.
Yes, you should increase them accordingly the load you database have. I'd get the timing of the redo switching and increase the size of the log members so the switching time is no less than 30 minutes.

Also, if the problem arises intermittently you should audit the applications or batch processes that might be generating a large amount of redo information. You can use log miner to see a clearer picture (stats) of what's filling your redo logs.