0

I have an GAE Java app now in development stage.

Starting today at 2013-07-22 13:33:33.452 PDT till 2013-07-22 14:00:18.678 (that's around 27 minutes), my app continuously got warmup requests in every 5~6 seconds, and so quota exceeded, and now it's not working now. As far as I know, we didn't do anything at that time and it just started getting warmup requests. (It's targeting just for Korean customers, and was early in the morning in Korea at that time, so I'm pretty sure that no one was using it.)

Most recent change I made to this app was 1) upgraded to 1.8.2 SDK, and 2) enabled dedicated memcache in app console's Application Settings page just to give it a try.

What's happening here? Is this a bug in SDK or dedicated memcache? Should I have to downgrade to SDK version 1.7.x?

ckang
  • 21
  • 3
  • 7
  • We got hit by the same problem on the python gae roughly 2 weeks before on some apps (even older ones). It was constantly issuing warmup-requests (and killing instances as it reached +30 idle instances). What helped us was setting the slider for idle instances back to automatic-automatic (it was 1 - automatic before). Never seen this since then. – T. Steinrücken Jul 23 '13 at 06:58
  • This is happening again just after the midnight, until 2013-07-23 05:15:24.579 PST. Some normal requests are still being processed, but full of warmup requests. – ckang Jul 24 '13 at 03:12
  • I've downgraded to 1.7.7, and let's see how it goes. – ckang Jul 24 '13 at 04:26
  • Now it's behaviour is a bit different. The "continuous warmup request" problem seems not happening any more. Beside that, my app has $4 daily budget (was $3 before) and usually only using around $1.5 a day. But after this incident, it's just using all $4 budget up in around 12 hours (about 10 hours today), and says it hit the budget limit. But the sum of all cost is just $1.14 right now. What's happening here? – ckang Jul 25 '13 at 01:41

0 Answers0