Hi, This is slightly OT, but I guess others have had the same concern so why not try ;-)
I've got several instances of Cocoon running behind a common httpd server via virtual hosts and mod_proxy. I'd like to make sure that no more than N simultaneous requests hit each of these Cocoon instances, to avoid the case where one instance gobbles up all the available connections. Ideally, requests over N should get a "500 - server too busy" responses, and they wouldn't hit the backend Cocoon instances. The MaxClients directive doesn't work for this IIUC, it's only usable at the global level. Has anyone got experiences to share? It looks like mod_tsunami would help, but we're running httpd 2.0.49, upgrading to 2.2.x soon, and it's for 1.3.x only. -Bertrand [1] http://bertrand.demiddelaer.org/mod_tsunami/