I have a Windows Azure applicaion (web-role) which I'm having serious problems with.
If there are problems with one of my databases which the application connects to (in this case, SQL Azure transient errors) many exceptions are thrown in my application. Every single exception thrown is caught and handled nicely, yet the IIS service eventually shuts down, or even worse, the entire web-role becomes un-responsive.
The number of exceptions is high, anything from 30-100 per second.
My application runs happily on 3 web-role instances, whilst experiencing this problem I have upgraded to 10 instances because I'm sure I read that caught exceptions are quite heavy on resources, but this made no difference.
There are no errors or (from what I can determine) helpful warnings in the event log.
.. and just to repeat, I can guarantee there are no uncaught exceptions occurring.
Does this sound like normal server behaviour?