Thanks Brian, I have optimized my code for the startup issue (carefully avoid JDO and initialization of classes involved in that framework). Also on startup, it usually flags the log with an info/warning message.
I looked at the log again. This time after browsing for several minutes (to make sure that warm up is likely not the culprit). The log has a lot of responses (1 in 10) in the 1-2 second time range for which the app stats reports only about 100 ms or so of the time. The rest of it is time spent in an unknown place I guess my point is that appstats captures only a part of the information needed to diagnose this response time. The rest of it is not available to the users of the App Engine. Can this information be made available beyond just the status page on the App Engine? For example, the status on the App Engine shows the static get latency close to 300ms on average. That still leaves me with particular requests that can be hung for an order of magnitude larger than the average. On May 18, 11:19 am, bFlood <[email protected]> wrote: > hi nicky, cool sight. > > those are most likely cold starts where the GAE infrastructure is > starting up a new handler for your request. unfortunately, there's not > a lot you can do about them on low traffic sights. it's actually worse > when you have a low traffic site that makes a lot of small, concurrent > requests (like map tiles). > > search the GAE groups and you'll find lots of discussions on cold > starts. the proposed remedy of reserved, "hot" instances will not > really help in the scenario either, sorry > > for out map tiles, I've just learned to live with it and make the > handler code for map tiles as efficient as possible. YMMV > > cheers > brian > > On May 18, 12:12 pm, Nicky <[email protected]> wrote: > > > > > > > I am wondering if there is explanation for this. I have a simple > > apphttp://dsm2bathymetry.appspot.comthatserves custom tiles for google > > maps. I turned on appstat on it and have noticed that there are large > > unexplained windows of time in the request that are not reported on. > > This gives me very little to go on. > > > For example this the appstat report on one such request > > 2010-05-18 11:04:10.899 > > 307 GET /tiles/-1751683366_tile334_790_11.png > > real=155ms cpu=0ms api=0ms overhead=3ms > > > The same request in the log > > 05-18 04:04AM 04.304 /tiles/-1751683366_tile334_790_11.png 302 6792ms > > 145cpu_ms 12api_cpu_ms 0kb Mozill.. > > > Approx 150ms are reported by appstat but what of the remaining > > 6792-150 approx 6 seconds of delay in the request. > > > This is just a sample, there are a number of requests for which it is > > unknown whats taking so long. The code for the app is available > > athttp://code.google.com/p/dsm2-grid-map/source/browse/#svn/trunk/dsm2-.... > > > -- > > 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 > > athttp://groups.google.com/group/google-appengine?hl=en. > > -- > 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 > athttp://groups.google.com/group/google-appengine?hl=en. -- 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.
