Hi Takashi, You wrote:
Actually, in that timeframe, I can see that only loading requests of > your application had higher latencies. > Those kind of hiccups are supposed to be automatically addressed and > fixed swiftly(almost always faster than manual intervention like me > doing right now). Additionally, you can set "Min Idle Instances" to a > certain level in order to mitigate this kind of risk, although it will > certainly cost you some more money. Yes, the loading requests have become very disruptive to my application. I am currently at Auto-Auto for the Idle Instances and Pending Latency settings. There is 1 client connecting to my application, and it provides a light load. I cannot see from my logs when an instance shuts down, but I can see when one is started. For example: 1. 2012-06-03 06:25:12.259 /xxx/sdtp?t=ACCESS_POINT_APP_DATA_UPLOAD 200 10716ms 0kb xxx.xxx.xxx.xxx - - [03/Jun/2012:05:25:12 -0700] "POST /xxx/sdtp?t=ACCESS_POINT_APP_DATA_UPLOAD HTTP/1.1" 200 64 - - "xxx.appspot.com" ms=10716 cpu_ms=5044 api_cpu_ms=708 cpm_usd=0.140171 loading_request=1 instance=00c61b117c977896b0d6d1161670c6b7518c43 <https://appengine.google.com/instances?app_id=s~sensorapp3&version_id=1-8.357553477529064533&key=00c61b117c977896b0d6d1161670c6b7518c43#00c61b117c977896b0d6d1161670c6b7518c43> 2. I2012-06-03 06:25:12.253 This request caused a new process to be started for your application, and thus caused your application code to be loaded for the first time. This request may thus take longer and use more CPU than a typical request for your application. 1. 2012-06-03 06:22:19.694 /xxx/sdtp?t=AP_RTC_SYNC 200 34ms 0kb xxx.xxx.xxx.xxx - - [03/Jun/2012:05:22:19 -0700] "POST /xxx/sdtp?t=AP_RTC_SYNC HTTP/1.1" 200 72 - - "xxx.appspot.com" ms=35 cpu_ms=38 api_cpu_ms=0 cpm_usd=0.001107 instance=00c61b117c3d078ba2eaebba354307e44c0a15 <https://appengine.google.com/instances?app_id=s~sensorapp3&version_id=1-8.357553477529064533&key=00c61b117c3d078ba2eaebba354307e44c0a15#00c61b117c3d078ba2eaebba354307e44c0a15> These two, back-to-back entries show something odd. The first request at 6:22:19 is handled by an existing instance, namely ...0a15. The next request comes in 3 seconds later. However, it is not handled by instance ...0a15. A new process is started, namely ...8c43. You are correct, when you say: Actually, in that timeframe, I can see that only loading requests of > your application had higher latencies. My question is: "Why was instance ...8c43 created to serve the request, rather than using ...0a15?" Cheers, David -- You received this message because you are subscribed to the Google Groups "Google App Engine" group. To view this discussion on the web visit https://groups.google.com/d/msg/google-appengine/-/eimEd2RRNIIJ. 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.
