We had a similar setup to AJs when I was at JD Edwards. Production instances, almost all 1 instance / physical server, were in the 200-300 range. Two DBAs.

We also helped support some development environments at times as well. We weren't slammed most of the time, but it was constant tinkering, monitoring, making sure that the stuff we'd set up to alert us ahead of time was working. And of course, something is always broken with 100s of instances.