Unfortunately, the problem is not yet solved. Once I activate a new
version, the latency is perfectly fine. But over time, things get
worse and painfully slow.

A couple of minutes after setting a new version to active:
Total number of instances       Average QPS*    Average Latency*        Average 
Memory
5 total (3 Resident)    1.083    81.4 ms         26.6 MBytes

2 hours later - the same version, the same code:
Total number of instances       Average QPS*    Average Latency*        Average 
Memory
21 total (3 Resident)   0.321    881.8 ms        30.8 MBytes

After less than a day, we have a latency of 1500 to 1800ms and our
monitoring tools start reporting failures that we can only fix by
manually activating a new version. Appstats says it's the CPU time ...
Almost every request is just a memcache lookup and in 99.9% not a
miss.

@Googlers
Do you have any idea why that happens? Is it due to your scheduler
(when things are slow, we do about 20 instances running)? Any hints?
Thoughts?

Best regards,
Fabian

-- 
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.

Reply via email to