We can feel the heat just by reading your requirement! Definitely worth considering more seriously. Added new issue "Support peak loads": http://restlet.tigris.org/issues/show_bug.cgi?id=198
Best regards, Jerome > -----Message d'origine----- > De : John D. Mitchell [mailto:[EMAIL PROTECTED] > Envoyé : mardi 14 novembre 2006 15:53 > À : [email protected] > Objet : Re: Re: Re: monitoring an application > > On 11/14/06, Piyush Purang <[EMAIL PROTECTED]> wrote: > > Yes, John this is a real world requirement this we are out for a > > coffee thingy :) > > Indeed. :-) > > Throttling usage during overloads is critically important to > gracefully surviving things like getting slashdotted. Doing this at > the application and container levels also helps protect against broken > services (i.e., the default page becomes the "out to lunch" page > instead of just falling over). > > Adding the ability to dynamically generate a useful, "static" page > (that can get aggressively cached at all layers) is even better in > terms of gracefully degrading service. E.g., look at how the big news > sites ended up having to do this manually during the 9/11 disaster -- > they created a simple, static page with very little graphics that they > updated whenever new information came along and they set the cache > time to something like 15 minutes. > > Have fun, > John
