estion a week or so ago (just search
> for pending_ms in the archives to find it).
>
> Waleed
>
>
>
>
>
>
>
> On Mon, Jun 27, 2011 at 7:20 AM, Spines wrote:
> > I'm using GAE/Java. I record the time it takes for my code to execute
> > inside the co
I'm quite happy with the new "scheduler knobs". My app takes 10
seconds on a loading request, and the current scheduler starts up new
instances too often, making too many users experience loading
requests. With the new scheduler, I'll be able to tweak the scheduler
so less users experience loadin
I'm using GAE/Java. I record the time it takes for my code to execute
inside the code and output it to the logs as a debug message. The time it
takes for my code to execute is usually just a fraction of the time that is
shown on the logs as the total time to respond to a request.
For example
@Ugorji
Shutting down an instance after 1 minute of inactivity would be bad
for java apps which usually take 5+ seconds to start up. The
scheduler will probably take into account a bunch of different
variables in determining how long to leave an instance running.
On May 20, 11:52 am, Ugorji wro
1. How will the "Always On" feature be handled?
On May 12, 12:14 am, Kenneth wrote:
> Greg mentioned he was putting together an FAQ so let's help him out!
>
> If you're going to answer this just put in your question into a single line,
> let's not try and answer them here or give opinions, there'
gt; than cpu.
>
> Under this proposed model, multiget/multiput are a single api call but the
> cost will be determined by the number of entities you are fetching or the
> number of entities and indexes you are writing with that single api call.
>
> Question for Spines:
> Why do y
Will we be charged for the time that an instance is up and running,
even if it isn't processing a request, or only for the time that it is
processing a request? Example: If an instance is running for a
minute, but only serves 1 request that takes 200ms, will we be charged
for 1 minute or for 200ms
I read that too. Seems like that is saying there is no free amount of
reserved instances.
On-demand Frontend Instances- 24 Instance Hours - $0.08 / hour
Reserved Frontend Instances -blank___- $0.05 / hour
Looks like the $9/mo gets you an SLA and the ability to be "infinitely
s
Where did you read that $9/month would get you one always-on instance?
On May 10, 1:54 pm, stephenp wrote:
> I believe for $9/month you get one always-on instance. Then, you need to go
> make your app thread-safe and turn on "multiple requests" for your app so
> each instance can handle more than
Similarly, with the new pricing, gets and puts will now cost the same?
I liked the old model that encouraged efficiency.
On May 10, 1:50 pm, Spines wrote:
> Doing a get of a 100 entities at once would be one API call right? So
> it would cost the same as doing a get of 1 entity?
-
Doing a get of a 100 entities at once would be one API call right? So
it would cost the same as doing a get of 1 entity?
--
You received this message because you are subscribed to the Google Groups
"Google App Engine" group.
To post to this group, send email to google-appengine@googlegroups.com.
I want to have at least 1 reserved instance so that my users will
experience less loading requests. Since we get 24 instance hours free
per day, could that be used on having a reserved instance? Or does it
only apply to on-demand instances?
Having 1 reserved instance cost $36 per month is very di
What will the memory limit of an app engine instance be in the future?
Is it going down? The backends are configurable from 128mb to
1024mb, but what is the limit for a regular instance?
--
You received this message because you are subscribed to the Google Groups
"Google App Engine" group.
To
Ikai,
The docs seem to indicate that master/slave is the way to go unless
you have a "mission critical" application. Maybe the docs should be
updated? Also when creating a new app, master/slave is the default
storage option, and you don't even see the High replication option
unless you click "edi
More info:
My cpu_ms for cold starts is consistently between 7 and 9 seconds, yet
the actual time it takes to respond fluctuates from 8 to 75 seconds.
Log lines:
2011-03-03 09:36:26.620 /?tab=recentedits&ajax=true 200 47085ms
8873cpu_ms 30api_cpu_ms
2011-03-03 09:25:38.135 /?tab=popular&ajax=true
For the past couple days, I've been experiencing huge cold start
times. They range from 10 seconds, all the way up to 75 seconds. On
average they seem to be about 30 seconds.
Just a couple weeks ago my cold start times ranged from 7 to 15
seconds. To make sure it wasn't due to any changes in my
The using the HR datastore page (http://code.google.com/appengine/docs/
python/datastore/hr/overview.html) says:
"This allows queries on a single guestbook to be strongly consistent,
but also limits changes to the guestbook to 1 write per second (the
supported limit for entity groups)".
I remembe
mance for entities only
> within entity groups or otherwise?
>
>
>
>
>
>
>
>
>
> On Wed, Mar 2, 2011 at 9:21 AM, Spines wrote:
> > Looking at the app engine status page, it seems that gets and queries on
> > the HR datastore are consistently faster than o
Looking at the app engine status page, it seems that gets and queries on the
HR datastore are consistently faster than on the Master/Slave datastore.
>From comparing the graphs, it looks like HR datastore queries are almost
twice as fast.
Why is this? Is this something we can expect to continue
I just had a loading request that lasted 37 seconds this time.
/ 200 37159ms 9480cpu_ms 30api_cpu_ms
--
You received this message because you are subscribed to the Google Groups
"Google App Engine" group.
To post to this group, send email to google-appengine@googlegroups.com.
To unsubscribe fro
Is the request time limit for loading requests still 30 seconds? I didn't
think there had been an increase, but I just had a loading request on my app
that completed successfully after 31652ms.
--
You received this message because you are subscribed to the Google Groups
"Google App Engine" gro
Just verifying - App Engine is hosted on servers located in the USA
right? And will stay that way?
The reason I ask is because I've been reading up about "fair use" for
using other people's images, and it seems that things are much clearer
legally if your servers are located in the USA.
--
You
Any idea when "Ability to reserve instances to reduce application
loading overhead" will be coming? Its been on the road map for about 6
months i think.
The reason I ask is that I'm considering using Grails for a new app
engine project. I deployed a sample application and the cold starts
were tak
It's most likely because you are experiencing a high percentage of
loading requests. If you are only accessing your app infrequently
then a large percentage of your requests will be loading requests
which take a lot of extra time. And this is probably what is bringing
your average up.
Once your s
I did some performance testing on the new high performance image
serving that uses getServingUrl.
If anyones interested I wrote about it here:
http://www.answercow.com/q/304002/app-engines-high-performance-image-serving-system
On Aug 18, 11:45 am, Spines wrote:
> Thank you App Engine Team
I'm not sure what you're talking about with respect to decoding the
Webservice, but my application uses a c# client to access it. Here is
the c# code where it accesses it:
HttpWebRequest req = (HttpWebRequest)WebRequest.Create(url);
req.Method = "GET";
req.KeepAlive = false;
using (var respstrea
getServingUrl works fine, returning a url of
http://localhost:8080/_ah/img/sxO56aAM_qVlARX4ym61bQ,
but when i navigate my browser to that url i get a security exception:
any ideas how to fix this, or is it a dev server bug?
Aug 18, 2010 8:14:45 PM com.google.apphosting.utils.jetty.JettyLogger
wa
Thank you App Engine Team! I had been struggling with how I would get
a decent solution for displaying lots of thumbnail images on one page,
and with this release and the getServingUrl function, it seems like my
problems are solved!
On Aug 17, 2:19 pm, "Ikai L (Google)" wrote:
> Hey guys,
>
> T
On the appspot.com admin control panel, I can't edit a datastore
entity if it has a ShortBlob property. When trying to edit it, the
shortblob is shown in an editable textbox as ��,�:3�. When trying to
save the entity, I get a page that says, "A server error has
occurred.". This happens when makin
I want to use 2 apps for my website. The first app would run the
website, and second app would be a bunch of random functions exposed
by an api.
The reason I don't want to lump it into one app is because I want to
cleanly separate the code of the random functions and the code that
runs the websit
I agree programmatic access to logs would be super useful. I opened
an issue specifically asking for this.
http://code.google.com/p/googleappengine/issues/detail?id=2953
. Please star it to increase its exposure.
--
You received this message because you are subscribed to the Google Groups
"G
Ahh, didn't think about disabling the indexing on all columns, thats a
good idea!
On Mar 11, 8:11 am, vivpuri wrote:
> I have experienced that Logger.info() logs dont always show up in the
> logs when viewed via Admin Console after few mins. Note, my app goes
> through lot of requests in short am
I'm wondering how reliable are the app engine logs? If I call
Logger.info("something"), and the call succeeds, will the entry
definitely show up in the logs? Or might it sometimes silently fail?
I have some data I have to persist with every request, but I will only
read the data once, about 30 mi
My cold start time for helloWorld is 1.5 seconds. If I add JDO to
that, it adds an additional 5-6 seconds for me on the first use of
JDO.
On Mar 6, 12:55 am, Francois MASUREL wrote:
> Here is what I get when I deploy my webapp with a very basic web.xml
> containing only one HelloWorldServelt, no
The idea of a cache is that at any time it could be out of date.
Trying to synchronize wouldn't be the intended purpose of a cache I
don't think.
--
You received this message because you are subscribed to the Google Groups
"Google App Engine" group.
To post to this group, send email to google-ap
If I could get a definitive answer on this that would be great.
I know that when your Google App Engine (GAE) app has 0 instances
running (because it has been idle for a bit), and a user requests a
page, the user has to wait for the instance to boot up and do all of
the instantiation which can cau
As long as there is at least one app instance running will that
guarantee another user won't get a loading request? My loading
requests take like 10 seconds to initialize Spring, Spring Security,
and JDO. I don't want any of my users to experience a 10 second
wait. The docs say that once there i
I know that when your app has 0 instances running (because it has been
idle for a bit), and a user requests a page, the user has to wait for
the instance to boot up and do all of the instantiation which can
cause the user to wait a significant amount of time.
My question is about the situation whe
38 matches
Mail list logo
