Has anyone managed to profile cold starts (I don't know whether that's even possible) to see where they take most of the time? Does the length of cold starts on GAE servers correspond to how long it takes on the development server?
I use GAE for Java so what I'll write below reflects that. I know the libraries I use employ annotations and that's got to have an impact. Jaxb (coupled with restlet) for example has got a terrible cold start time. My warmup request does exactly that, it triggers a jaxb warmup. But lately it has been failing. Since that seems to affect mostly the warmup request, users don't seem to be affected terribly (we have very low traffic at this stage though). I wonder whether the Google Engineers could implement something similar to what Android does: there's a resident Dalvik VM and all new processes fork from it. Our prototype process could have the jvm, the web server and all the rest that's common. Just a thought. E -- You received this message because you are subscribed to the Google Groups "Google App Engine" group. To post to this group, send email to [email protected]. To unsubscribe from this group, send email to [email protected]. For more options, visit this group at http://groups.google.com/group/google-appengine?hl=en.
