WebPagetest Forums - All Forums

Passive performance monitoring - How to measure Base page load time?

Wed, 01 Sep 2010 17:39:24 +0000

Since early this year ive been working on and off on an analytics script for personal use (may make it commercial service .. dunno) . Due to time constraints i haven't been coding for last 2 - 3 months, but now am ready to start work on it again.

Patrick's post was a major encouragement re-affirming that there is a market for such services.

But the question is how can one really measure the performance?

My current method is very lame... i.e. it saves the time in a javascript variable, compares that to the time after onLoad and reports that back. This does not account server delays for the base page.

Boomerang by Yahoo helps solve this by setting timestamps in cookies and checks that on the next pageload, but this helps measure basepage time only if there is > 1 pageview/session. The bounced sessions wont be measured.

Just wondering if there is a more elegant way to monitor the complete pageload time? Anyone have any interesting tricks up their sleeves?

I promise to give free beta access (provided I have something launchable) to people who can point me in the right direction

problem with getwork.php

Wed, 01 Sep 2010 15:40:22 +0000

Hi,

I have installed webpagetest v1.9 on Apache+Tomcat and excuted urlBlast on XP.

I can see hosted webpagetest (host XX.XX.XX.218) site. but when I run urlBlast (client XX.XX.XX.216) following error occurred on apache.
------
[Thu Sep 02 00:01:59 2010] [error] [client XX.XX.XX.216] PHP Warning: mkdir(): Permission denied in /usr/local/tomcat6/webapps/ROOT/wpt/content/work/getwork.php on line 28
[Thu Sep 02 00:01:59 2010] [error] [client XX.XX.XX.216] PHP Warning: fopen(./times/Test.tm): failed to open stream: No such file or directory in /usr/local/tomcat6/webapps/ROOT/wpt/content/work/getwork.php on line 29
[Thu Sep 02 00:01:59 2010] [error] [client XX.XX.XX.216] PHP Warning: fopen(/usr/local/tomcat6/webapps/ROOT/wpt/content/work/test/lock.dat): failed to open stream: No such file or directory in /usr/local/tomcat6/webapps/ROOT/wpt/content/work/getwork.php on line 34
----

My urlblast.ini like this:
---
[Configuration]

Test Type=4
Startup Delay=30
Log File=D:\03_Performance\Front-End\Web Site\log
Thread Count=1
Timeout=60
Reboot Interval=720
Clear Cache Interval=30
browser=pagetest.exe

; Where to get work from
Url Files Url=http://XX.XX.XX.218/work/
Location=Test
Location Key=qa4
---

and my location.ini like this
---
[locations]
1=Test_loc
default=Test_loc

[Test_loc]
1=Test
default=Test
label=Test Location

[Test]
browser=IE 7
latency=0
label="Test Location 1"
localDir=/usr/local/tomcat6/webapps/ROOT/wpt/content/work/test
key=qa4
---

I guess this is permission problem with httpd.conf
But i don't know exactly what i do.

Any help on this would be gratefully appreciated.
Thanks always for great tool and you.

9 seconds for DNS look up, how to minimise, what is the standard DNS time?

Sun, 29 Aug 2010 23:39:30 +0000

Hello Folks,
I have checked my Google webmaster tool and found my page load speed is very slow, so i checked with Webpagetest and it showed me this http://www.webpagetest.org/result/100829...1/details/ more than 9 seconds of DNS look up and my other website http://www.webpagetest.org/result/100829...1/details/ is also taking around 1 second for DNS how can i minimize this, what is the procedure and what is the standard DNS look-up time for an ordinary website?
Thanks in advance.

Why are some resources being downloaded twice?

Sat, 28 Aug 2010 15:38:04 +0000

Been referring to this site for ages, and thought I should probably make an account and join in the fun!

Test result: http://www.webpagetest.org/result/100828_3R2D/

I have used some CSS sprites, particularly one called backgrounds.png. I have used the two images in that sprite once each, using the following CSS.

#mainheader, #background {
background:#000 url('/media/img/sprites/backgrounds.png') no-repeat top left;
}

#mainheader {
background-position: 0 -643px;
}

#background {
background-position: 0 0;
}

Why does IE download backgrounds.png twice? I also have the file PIE.htc being downloaded twice...

(Any other optimisations are welcome!)

Ah, seems that CSS3 PIE causes this problem on elements that it's applied to. Useful to know. Still, any other optimisations are welcome!

How can I translate what all this means?

Fri, 27 Aug 2010 19:46:28 +0000

I am hosting a DNN Portal at PowerDnn and the site is loading rather slowly. One of the tips in the knowledge base is to come here and do a speed test. I did that and it's taking the site 5.8 seconds to load. My larger problem which I am working with them to solve is that on the initial load in the morning, the site is taking nearly 53 seconds to load.

PowerDNN is known to be good at DNN Hosting and are supposed to be the best in the business at what they do. Can anyone tell me by running the report on this URL: http://www.provsnusa.com.dnnmax
what is going on and why it's taking so long to load?

I'm wondering if it's a custom module I loaded or the Flex 2 skins... Not sure how to translate all this data.

Thanks ahead,

Dan

IE6

Fri, 27 Aug 2010 08:10:49 +0000

Pat

I seem to remember some IE6 agents at one time (or I might be thinking of a locally hosted setup?)

I'm assuming that there are none either because it's a pain or because IE7 is close enough. Could you confirm?

(IE6 is a must for our target markets so all other things being equal it would be a nice to have - especially in China and any other asian agents locations that might be coming along. Don't suppose you have any plans for an Indian agent? That would be awesome.)

BR as always

Neil

How can I fix "(No max-age or expires)" with Amazon CloudFront?

Fri, 27 Aug 2010 01:48:49 +0000

So here's the deal. I'm using Amazon's CloudFront product to deliver the .gif files for my site's template. There are about 30 individual files so it should help speed up the site. At some point I might get around to making a CSS sprite, but for now I'm not worried about that.

The problem now is when I look at my WPT results, it shows many lines of:
FAILED (No max-age or expires) - http://example.com/example.gif (one for each gif file).

So I know how to set the expires on my regular web server. I use apache so I put this in my .htaccess in the root:

Code:

# 2,592,000 is 30 days caching

Header set Cache-Control "max-age=2592000, public"

So the main question here is how do I enable that sort of cache control on AWS? I'm using a subdomain cdn.example.com as the CNAME for AWS, however the *domain* name of my main site is example1.com and the CDN *domain* name is actually a different domain, e.g. cdn.example2.net.

(The reason I'm using a .net domain for the CDN files is that domain is cookieless which is supposed to be a good thing when it comes to speeding things up)

flushing response performance gains?

Tue, 24 Aug 2010 18:04:02 +0000

I play with Django web framework for most web stuff, and the default template system doesnt support flushing.

As in since its a MVC framework, all computations are done before it starts to parse the templates...

Say I have a page that takes 1 seconds to generate on the server. The browser is doing nothing during that 1 second.

If I can somehow flush this portion instantly

Code:

And the rest after all the computations, DB queries, etc are completed.... does this mean that during that 1 second wait time, the stylesheet and javascript is already downloaded and will not block anymore once the rest has arrived?

I have some javascript and css which needs to be used on all pages...

to be more specific.... will it work if i have the

Code:

after the flush?

Page Speed and AdSense

Tue, 24 Aug 2010 05:41:05 +0000

Hey Everyone,

I am trying to avoid inline JavaScript as much as I can. I am also trying to have all my JavaScript execute after the onload event takes place. I know Google stated in their TOS that publishers are not allowed to modify their code. I have a few questions about this:

#1 - Can Google Adsense be loaded asynchronously without violating their TOS? When it is synchronous and inline, it may delay the body onload event from occurring.

#2 - Can Google Adsense be placed into a function that is called after the body onload event takes place without violating their TOS?

#3 - Can Google Adsense JavaScript be placed into an external JS file without violating their TOS?

#4 - Can Google Adsense be placed into a frame (with no content) as long as the frame is part of a bigger page (with content) without violating their TOS? I am thinking the frame src value could be set once the body onload event occurs.

#5 - How do you increase your page speed when dealing with ad systems such as Google without violating any rules?

Thanks in advance for any suggestions and / or information.

Sincerely,
Travis Walters

Video ON > Start Render green line is off

Sat, 21 Aug 2010 12:09:24 +0000

I found a bug having to do with the green Start Render line in the waterfall charts.

I ran a few test on page A, without checking the "Create Video" checkbox.
Results:
http://www.webpagetest.org/result/100820...acda76aea/

Start Render line is drawn right after the CSS file has finished loading and this is as expected/correct.

Compare this to the waterfall charts for the other test runs, where I checked the "Create Video" checkbox.
Results: http://www.webpagetest.org/result/100820...b3d1062a8/

The green Start Render line is drawn too early, at the Connect time for the HTML. This of course is wrong since the browser has not received any bytes...

Pat > can you fix this? Would be great!

It's normal to have so different test results ???

Fri, 20 Aug 2010 16:25:45 +0000

Hello,

I have made some consecutive test on my site and I receive very different and confusing results.

1.http://www.webpagetest.org/result/100820_3A87/
2.http://www.webpagetest.org/result/100820_3A8A/
3.http://www.webpagetest.org/result/100820_3A8D/
4.http://www.webpagetest.org/result/100820_3A8J/
5.http://www.webpagetest.org/result/100820_3A8N/

Beside the very different Start Render time are other very strange things.
1.Why somethimes the Reapeat View is slower than the Firs View (see test 1 and 3)
2.Why for test 4 the reqests are paralellized 2 by 2 and for others they aren't?

It is because the hosting server or because my site code ???

Allowed Frequency of Automated Tests

Thu, 19 Aug 2010 19:01:43 +0000

I have 8 URLs I'd like to use in an automated test by calling http://www.webpagetest.org/runtest.php.

I have a process that can hit each URL at a scheduled time. If I were to schedule to have each URL hit weekly, would it be an issue?

Is there a point at which webpagetest.org begins to care about someone hogging resources? I don't think I'd ever need to go more frequently than weekly, but I may want to add additional URLs at some point.

Thanks.

Embedded Youtube vids - weird, large Content Download bars

Mon, 16 Aug 2010 10:46:22 +0000

hi Pat,

check out http://www.micazu.nl/vakantiehuizen/rosa...rum-6159/.
It's a page on a travel site with info about a rental house.
Scroll down and you see the 2 embedded Youtube vids. Nothing special, you see this all the time on websites.

I ran the page 3 times through WPT - IE7 from location Amsterdam - , results here: http://www.webpagetest.org/result/100816...f0854e5e3/

See the huge blue bars in the First View waterfall charts? Content download time is high for both YT objects. The response header specifies the Content-Length and that's about 1 KB. And, Bytes In (downloaded) is 1.5 KB.
Clearly - and this makes sense - this is just the video *player*, not the actual video. Further down in the waterfall there are two objects served from ytimg.com which are Youtube SWF files. These are ~140 KB and show normal Content Download times.

So, I don't understand the long blue bars for those first youtube.com objects.
I did some investigation and tried to find out why this is.

1) Is the Amsterdam location behaving weird?
Answer: No.
View Paris (IE7) location test results here: http://www.webpagetest.org/result/100816_301S/
The Content Download times are very high here too.

2) Is it an IE7 thing?
Answer: No.
View Dulles IE8 results here: http://www.webpagetest.org/result/100816_303J/
Blue bars are not was long but they are still longer than expected (maybe because Dulles is better served by YT CDN than the poor little servers in Europe ;-) )

2) Is it a Webpagetest.org thing (versus IE locally)?
Answer: I think so.

I visited the same page on micazu.nl with my local IE7 on Vista and used Pagetest 2.0.0.215 to see what's happening.
The waterfall showed no weird long blue bars. They were all normal. The response headers were normal too and similar to those on the wpt.org tests.

Note: in the Amsterdam, Paris and Dulles test result you can see the high CPU utilization at the time of the 'long blue bars'...

More info:

I actually did the tests from Amsterdam twice and only posted the link for the 2nd batch in the first post in this thread.
So here's the link to the 1st batch: http://www.webpagetest.org/result/100816...87c63734b/

The first test in the first batch does *not* show the long blue bars !
http://www.webpagetest.org/result/100816...1/details/

split css file to improve perceived latency?

Sat, 14 Aug 2010 14:13:29 +0000

For some unknown reason, the css file im working with is about 97 KB (un gzipped) ... I know thats horrible coding and will be completely redone later on...

For now, Im curious if i split the css file into 2 smaller chunks, will it improve the start of rendering? As what I understand, the browser wont start rendering the page until all the CSS is downloaded... so 2 files would be downloaded in parallel? If i combine this with domain sharding, this should result in faster output i guess...

For the final phase (includes rewrite of the css) my plan is to use 2 css files on all pages... one would be the base stylesheet which contains elements used on almost all pages and another stylesheet containing elements specific to the type of page (homepage, product page, listing, cart, account, etc, whatever).... and using preloading to make the second page visited to be almost instant.

What do you think? Bundle everything into one css or split it?

Auto Spriting?

Fri, 13 Aug 2010 18:22:28 +0000

Anyone know of any tools to generate css sprites automagically with very less input from the developers?

Currently I am playing with SmartSprites, but would like to know better alternatives. This tool kinda makes things look different and need to hack a lot to make everything look fine... not a nice approach.

I am building a set of tools which can be integrated into the build/deploy procedure of the project, so the developers play with the raw unoptimized code, but the deployed version makes the sprites, bundles javascripts, do some preloading, etc.

Looking for some pointers for spriting.... Are there any other tools i should look at? The project is in python, but we can use any tool which runs on linux command line.

We were looking at using data URI (+ MTHML) , but decided against it since it made the CSS files really bloated.

CSS @font-face and IE 7 & 8 404 errors

Fri, 13 Aug 2010 07:40:55 +0000

Whats the best way to handle embedded hosted fonts? IE requires an .eot font and doesn't know how to handle .woff and .otf so it eventually throws a 404 after the .eot loads.

You can see the issue in my test results:
http://www.webpagetest.org/result/100813_2VR8/

Chrome, Firefox and Safari will load the .woff and skip over the .eot without making a request.

I have been contemplating serving a specific ie style sheet with only the .eot fonts but I hate the idea of browser specific style sheets. I guess I shouldn't be complaining with first load times under 1 sec on WordPress but I am a performance junkie.

Proper cookie usage

Thu, 12 Aug 2010 09:45:50 +0000

Hi,
i want to ask what do I have to do to have checked "Proper cookie usage". Is there something in htaccess (for images/scripts/styles..) I should add?

GZIP js & css

Thu, 12 Aug 2010 07:25:58 +0000

In the results of webpagetest the results of gzip seem to be okay.
But Google Webmaster tools say, that I should gzip js and css also.
Should I?

How is the "Fully Loaded" time measured?

Thu, 12 Aug 2010 00:47:46 +0000

Hi,

From what I understand, time to "Document Complete" is measured when the onLoad event in the browser gets fired, correct? I'm curious about how the "Fully Loaded" time is calculated. Is there a similar browser event for this?

Thanks!

What is easy to do to improve speed from test result?

Wed, 11 Aug 2010 23:42:06 +0000

http://www.webpagetest.org/result/100811...imization/

This is my performance review but I dont really know what to do next or what I would be able to improve easily with minimal knowledge...

I am using wordpress plugin W3 Total Cache by W3 EDGE ® to improve load speed (the site loads very slow) but I think there are things I have to do manually. for example I have an option to load js & css scripts to minify. Would I load the same files that the review says I should compress?

If I can get some suggestions and explanation how to optimize my site I would appreciate it

Thank u to whoever can help