AppStats would not measure SSL overhead as it starts tracing requests after SSL termination. From your screenshot I see no time spent in executing RPCs, so probably these requests are consuming high CPU cycles. Do you see anything in cpu_ms in request logs? If so, have you tried increasing frontend instance class and see if that has any impact on latency?
On Saturday, 15 November 2014 23:23:48 UTC+1, Alex Fu wrote: > > I recently turned on SSL for my AppEngine (Java) project and noticed a > decrease in response time. I enabled AppStats for AppEngine and proceeded > to hit an endpoint of mine: Screenshot > <https://lh4.googleusercontent.com/-qUv7q44ED3M/VGfRF_c5QJI/AAAAAAAACIQ/yi3RsltYXMI/s1600/Screen%2BShot%2B2014-11-15%2Bat%2B5.16.11%2BPM.png> > > AppStats reports zero overhead (real=5123ms api=0ms overhead=0ms cost=0 > billed_ops=[]). > I know SSL can introduce some overhead and latencies but a ~5s wait is a > bit much for a simple GET that normally take less than a second to complete. > > Is there any better way to debug this? > -- You received this message because you are subscribed to the Google Groups "Google App Engine" group. To unsubscribe from this group and stop receiving emails from it, send an email to google-appengine+unsubscr...@googlegroups.com. To post to this group, send email to google-appengine@googlegroups.com. Visit this group at http://groups.google.com/group/google-appengine. For more options, visit https://groups.google.com/d/optout.
