App engine problems

[conman]

Hello,

recently I am observing *many* DealineExeededErrors in my app!

This happens even when serving only static content without any
datastore operations.
I really think it's not my app because 1 h ago it worked quite normal.

This sucks because we just send out customer information and now our
site is not stable!
I guess this is what beta is about...

Anybody else observing this problems?

Regards,
Constantin

[service G2100]

My application use 0 datastore connection. And it works just fine.

I would recommend you check on your script. I got experenice with wrong script and hit the high cpu meter.

Just a clue for you.

Regards.

2009/2/6 conman <constantin...@googlemail.com>

[johnP]

I noticed this with my app yesterday in the afternoon - a bevy of
deadline exceeded errors, and an increase in zipimport events in my
logfile (using appengine patch). Later, it got better...

johnP



On Feb 6, 7:31 am, conman <constantin.christm...@googlemail.com>
wrote:

[Devel63]

I saw the same thing yesterday late morning, it was great last night,
and again bad this morning. Among other things, a page which normally
takes <70ms times out. Interestingly, I see that the while the time
was >9s, the ms-cpu was <1s ... which seems weird to me. Usually, I
see actual time < cpu time.

These pages are not doing any DB writes, but do write to memcache and
sometimes read from the DB.

[Marzia Niccolai]

Hi,

There have been periods this morning with higher than average serving
latency, but so far this is all we have seen:

http://code.google.com/status/appengine/detail/serving/2009/02/06#ae-trust-detail-helloworld-get-latency

-Marzia

[johnP]

"periods this morning with higher than average serving latency" are
displayed to our customers on a big white screen that says SERVER
ERROR in a big bold font. Just FYI.

:)

On Feb 6, 9:16 am, Marzia Niccolai <ma...@google.com> wrote:
> Hi,
>
> There have been periods this morning with higher than average serving
> latency, but so far this is all we have seen:
>

> http://code.google.com/status/appengine/detail/serving/2009/02/06#ae-...
>
> -Marzia

[johnP]

Actually, I don't have any customers yet (using the app) - but that's
just a quibble. :)

[conman]

Hi Marzia,

> There have been periods this morning with higher than average serving
> latency, but so far this is all we have seen:

I am pretty sure there was something wrong. I am sitting in Europe,
maybe that's the reason...

But I tested my site multiple times with
http://www.hyperspin.com/
which tests site accessibility from multiple locations round the
world.

Around 8:00 am PST it reported 500 errors from everywhere around the
world, sometimes one or two locations worked ok but the next time I
tried I got an error again.

Now I checked our site again and everything seems quite normal.

It's making me nervous that we (or at least I) are experiencing
*major* problems and there seems to be no explanation for this.

Maybe someone can look into that to make sure it doesn't happen again?

Regards,
Constantin

[Geoffrey Spear]

On Feb 6, 12:27 pm, johnP <j...@thinkwave.com> wrote:
> "periods this morning with higher than average serving latency" are
> displayed to our customers on a big white screen that says SERVER
> ERROR in a big bold font.  Just FYI.

So design your app to not do that?

[johnP]

Yes - a spiffy, cheerful, and optimistic 500 message will be in-place
for launch.

[conman]

This is not the 500 page of our app. The reported problems resulted in
a 500 page of the google service.
So actually we cannot 'design the app to not do that' :)

Regards,
Constantin

[conman]

The error report on
http://code.google.com/status/appengine/detail/serving/2009/02/06#ae-trust-detail-helloworld-get-latency
from yesterday 11:00 am talks about a memcache problem which 'didn't
affect the perfomance or uptime of applications'

I hope this problem was the reason for the heavy error rate of our
application but then this statement is not quite true!

But most importantly the problem seems to be solved and everything is
back to normal :)

Regards,
Constantin

[Devel63]

Again this morning. Here is an interesting log report, showing the
Django loading issue (note that it times out without really taking
much cpu resources)

1.
02-08 12:03PM 01.519 /hint/request 500 9025ms 672ms-cpu 1kb
Mozilla/5.0 (Windows; U; Windows NT 6.0; en-US; rv:1.9.0.6) Gecko/
2009011913 Firefox/3.0.6,gzip(gfe)
See details

E 02-08 12:03PM 10.534

<class 'google.appengine.runtime.DeadlineExceededError'>:
Traceback (most recent call last):
File "/base/data/home/apps/judysapps/5.331236991686126139/
main.py", line 64, in <module>
main()
File "/base/data/home/apps/judysapps/5.331236991686126139/
main.py", line 61, in main
util.run_wsgi_app(application)
File "/base/python_lib/versions/1/google/appengine/ext/webapp/
util.py", line 76, in run_wsgi_app
result = application(env, _start_response)
File "/base/python_lib/versions/1/django/core/handlers/
wsgi.py", line 189, in __call__
response = self.get_response(request)
File "/base/python_lib/versions/1/django/core/handlers/
base.py", line 110, in get_response
from django.views import debug
File "/base/python_lib/versions/1/django/views/debug.py", line
3, in <module>
from django.utils.html import escape
File "/base/python_lib/versions/1/django/utils/html.py", line
1, in <module>
"HTML utilities suitable for global use."



On Feb 7, 1:58 am, conman <constantin.christm...@googlemail.com>
wrote:
> The error report onhttp://code.google.com/status/appengine/detail/serving/2009/02/06#ae-...