Huy, > Let's say I have webapps A and B, and A is more critical webapp. Let's > say my connector's pool size is 100, and there are 100 concurrent > requests destined for A and B each (so total 200 requests here). I > would like to allocate 70 threads to process A's requests, and only 30 > for B's.
You could do this by fronting Tomcat with a web server like Apache httpd or a proxy server, and just use two different connectors. I think this would be a more elegant solution than trying to hack-together a short-circuiting, load-balancing RequestProcessor (probably a Filter would be better, but not by much). -chris
signature.asc
Description: OpenPGP digital signature
