Just as an experiment we ran this job again, this time using JMeter to "prime" the VM's at a sustained 25 requests per second for 15-20 mins before the tasks were queued up.
No impact - same request aborted rate as before. So it doesn't seem like it's a problem with availability of "warm" vm's... J On Aug 27, 10:08 am, "Jan Z/ Hapara" <[email protected]> wrote: > Had another data run complete just now - a url-fetch focused job > (talking to Google services only, no rpc, no db, little memcache) with > 2,500 tasks spread across 5 queues at 6/m each. Average task is 4-8 > urlfetches, runs in 17s, with < 100msec CPU time. There was minimal > to no other load on the app at the time. > > Job completion: 6h 24m > > Effective task run rate (taking the "requestaborted..." + requeue > into account): 384 per hour. Expected run rate: 1,800. > > Effective task queue efficiency: 21%. > > Something is amiss here... According to the docs, this is way below > any published quotas, we ought to be able to run this off a single > queue at 20/s? > > Some insight from Google around what to even look for would be hugely > appreciated!! > > J -- 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.
