Please Stop The DeadlineExceededError's

171 views
Skip to first unread message

Kaan Soral

unread,
Sep 23, 2014, 1:18:02 PM9/23/14
to google-a...@googlegroups.com
At first it was tolerable, now it's just frustrating and painful (after months)

I think it comes and goes in waves, hits multiple services, causes them to exceed the Deadline's, happens everyday, fills the logs, probably happens with a very small but significant probability

Can't services retry on their own, handle this issue?

Is anyone else experiencing this issue, irritated by it?
I remember discussions related to the issue when it first started, I assumed it's one of those issues that comes, gets fixed and goes, but this one stayed ...

Qian Qiao

unread,
Sep 23, 2014, 6:31:38 PM9/23/14
to google-a...@googlegroups.com

Kaan Soral

unread,
Sep 23, 2014, 8:38:51 PM9/23/14
to google-a...@googlegroups.com
Thanks for the link, interesting article / very old, however I found the logging advice interesting

I log heavily, especially some requests cause maybe 1000 logs, would this cause DeadlineExceededError's for taskqueue's/other services?

I was always under the impression that the logging system is lightweight, is this not so?

Nick

unread,
Sep 24, 2014, 7:44:35 AM9/24/14
to google-a...@googlegroups.com
I would guess that 1000 log requests per request is highly significant. Your best bet is to adjust your log level, maybe up to warning or error, deploy and see if there is any noticeable change.

You haven't quantified how often you get Deadline Exceeded, so it will be hard for anyone else to answer your question. Apart from instability with internal services, exceeding deadline is an issue with the code you deploy. Appengine has generally been getting more stable over the last few months, and now issues are detected and generally sent around on the notification group in a much more proactive fashion.

If you are getting this at a level that is driving you to distraction, then it is either:
a) caused by you
b) caused by data centre configuration that you suffer more than others (such as networking configuration to a particular geo location or unusual use of a service)

You can try making quantifiable change to resolve (a) 
You can raise a support ticket to resolve (b) - if you have support.
Reply all
Reply to author
Forward
0 new messages