mod_ruby with Apache 1.3

I am being hosted by a system with the following specs:

RedHat 9
Apache/1.3.29 (Unix)
mod_ruby/1.0.7 (recently upgraded to 1.1.1, but that didn't help)
Ruby/1.8.1

On several occasions, an apache process "gets stuck", insofar as it eats
> 95% CPU, and never dies.  Apache as a whole can "carry on", serving other HTML requests, etc.

The circumstances under which this problem occurs remains a mystery that
needs further and continued investigation, but so far, I have made the
following uncertain observations:

- It seems to only occur when/after an rhtml page is requested by a browser.
- The rhtml pages in question contain eRuby markup
- It has occurred with rhtml pages whose eRuby code is solely 
ERuby::import statements
- It might be the case that the problem occurs on certain special 500
errors.
- It is not, as yet, reproducible at will.
- I have personally witnessed instances where my browser will make the
HTTP connection, but otherwise hang, with no page downloading.  A
"stop"/abort in the browser and retry fixes the problem from the user 
side.  I do not know if this causes the hang.

I will continue to watch carefully and try to characterize the problem
in better detail, but I was wondering if anyone else has encountered
this issue or anything similar?  If so, are there any resolutions?

Hoping I don't have to go back to PHP,
Alex