More and more user facing requests are going into server instance startup
and getting served.
Please tell us what is good settings to avoid this.
-Aswath
On Mon, Jan 27, 2014 at 11:15 AM, Aswath Satrasala
aswath.satras...@gmail.com wrote:
There is no traffic. There is one instance running.
You should look at your timestamps, some of your warmup requests are
failing (DEE) and then you have activity that can't be served by a running
instance, so another instance is being started to serve that request.
Can you do anything to speed up your startup times ? Does it make sense to
Currently latency is set to:
Min Latency : 7.5
Max Latency : 14.5.
-Aswath
On Wed, Jan 29, 2014 at 4:05 PM, timh zutes...@gmail.com wrote:
You should look at your timestamps, some of your warmup requests are
failing (DEE) and then you have activity that can't be served by a running
There is no traffic. There is one instance running. I issued a new
request. A new instance is started and served.
Appengine team, please help us to avoid this or provide your guidelines, so
that a new instance is not started for a user facing request
Thanks
-Aswath
On Fri, Jan 10, 2014 at
All I can suggest is starring Issue 7865 (User-facing requests should never
be locked to cold instance starts) at
http://code.google.com/p/googleappengine/issues/detail?id=7865.
On Friday, 10 January 2014 06:23:09 UTC, aswath wrote:
Hello,
The user facing requests are going into the
You do have warmup requests enabled ?
On Friday, January 10, 2014 2:23:09 PM UTC+8, aswath wrote:
Hello,
The user facing requests are going into the instance startup and getting
served. Any settings recommended recently for avoiding this.
My settings are
instance_class: F2
Yes, it is enabled.
On Fri, Jan 10, 2014 at 3:54 PM, timh zutes...@gmail.com wrote:
You do have warmup requests enabled ?
On Friday, January 10, 2014 2:23:09 PM UTC+8, aswath wrote:
Hello,
The user facing requests are going into the instance startup and getting
served. Any settings
On Friday, January 10, 2014 7:58:54 AM UTC-6, Ian Marshall wrote:
You could increase your pending latency settings to the maximum allowed,
in order to try to have your app wait as long as possible for an existing
app instance to serve a web request. I use:
automatic-scaling