That's the thing about exponential backoffs: it grows so quickly. It goes
from 1-2-4-8-16-32 seconds. That's only 6 HTTP requests, and if you set it
to 8 seconds, it'd be only 3 HTTP requests worth of savings. I certainly
hope you're not deploying via 3G, as the overhead of 3-6 HTTP requests is
tin
Whenever I deploy new version/update to my GAE applications (be it
Java or Python--it does not matter) it never happened to me that new
version was ready earlier than 8 seconds (mostly 32, 64 or even
later). I do not believe it is just me. Can I somehow configure on my
side when to check for readin