My backed is triggered every 90 minutes from a cron job, and usually takes
between 10 and 20 minutes to do its work.  Since April 26th, it seems to be
taking 30 minutes at best, and ~100 minutes at the worst case.  On average,
its about a 5x increase in run-time.  So, I investigated to try to find out
which API calls are contributing the most to this increase.

It seems that its a call to taskqueue.add() that's gotten really slow.  So,
I ran a smaller version of my backend task on a front-end instance first,
and then again on the backend instance to try to compare run-times.  And
yes - the exact same HTTP request took 406 seconds on the backend, and only
7.5 seconds on the front-end instance.

This particular HTTP request leases a bunch of tasks from a pull queue, and
then processes them and re-inserts them back into the same pul queue.  I
put logging.debug() calls before and after taskqueue.add(), and noticed
that on the front-end that API call takes on average 8ms, while on the
backend it takes on average 600ms (around 2 seconds, worst-case).

This behavior has been happening since April 26th.  Anyone else seen this?

Regards
- Rishi

-- 
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