wesleyhales / speedgun Goto Github PK

During report run, there are two methods that post either perf data or base 64. These are not in sync and should ideally be ran as one.

Speedgun.io + CoreOS hosted environment will occasionally crash without any significant log messages. When I ssh into the instance and docker ps the Status shows "Up Less than a second" and then the containers start to fire up.
Definitely a bug, because just ssh'ing and running docker ps seems to restart the containers.

I just implemented a basic crawler here:
https://github.com/wesleyhales/speedgun/tree/spider

• needs refactor/cleanup
• needs to be integrated with new CLI argument

$ vagrant up
Bringing machine 'default' up with 'virtualbox' provider...
==> default: Checking if box 'chef/centos-6.5' is up to date...
==> default: Clearing any previously set network interfaces...
==> default: Preparing network interfaces based on configuration...
    default: Adapter 1: nat
==> default: Forwarding ports...
    default: 8080 => 8081 (adapter 1)
    default: 22 => 2222 (adapter 1)
==> default: Running 'pre-boot' VM customizations...
==> default: Booting VM...
==> default: Waiting for machine to boot. This may take a few minutes...
    default: SSH address: 127.0.0.1:2222
    default: SSH username: vagrant
    default: SSH auth method: private key
    default: Warning: Connection timeout. Retrying...
==> default: Machine booted and ready!
==> default: Checking for guest additions in VM...
==> default: Mounting shared folders...
    default: /vagrant => /Users/---/development/src/speedgun/server
Failed to mount folders in Linux guest. This is usually because
the "vboxsf" file system is not available. Please verify that
the guest additions are properly installed in the guest and
can work properly. The command attempted was:

mount -t vboxsf -o uid=`id -u vagrant`,gid=`getent group vagrant | cut -d: -f3` vagrant /vagrant
mount -t vboxsf -o uid=`id -u vagrant`,gid=`id -g vagrant` vagrant /vagrant

Installing the vbox extensions on the centos machine doesn't help.

Both could stick with Vagrant for consistency. The direct Docker way should be reserved for advanced users.

When the report is running on speedgun.io, the UI polls for base64 data coming from phantom. The angular service hits http://speedgun.io/rest/performance/checkimage?uuid=??? every few seconds waiting for the next image. This data should be stored in localStorage and reused if the person who runs the report revisits speedgun.io at a later time.

Currently base64 images/screenshots or only temporarily stored per report.

Just because they submit an email, doesn't mean I have consent to show :)

Currently I am abusing the material CSS classes to get what I want out of them.
We need:

• Better text layout
• a modal should pop up on URL errors

This is a performance site... need to dumb down the bg image or rework the style/them all together.

Just reporting ;)

currently the report returns too much data. We should only be showing the stuff we calculate from api start and end times.

I've tested for the main newspaper in Sweden: http://www.aftonbladet.se

And I didn't get any page load time:
http://speedgun.io/#?uuid=c292df23-3cc2-4d95-b243-ed66561ca5fb

When I test in Chrome the loadEventEnd is fired. Could be a bug somewhere?

Best
Peter

I think like this:
The most important metrics first, making it easier for people to see the important things. Then it's just the hard part defining what's most important :)

I like that page load and perceived are in top, but then I would move up slowest and fastest assets. Then I would hide things like Unload Event Start, Redirect start, navEvents etc and make it possible to show them only if you want.

Best
Peter

When I'm testing http://www.cybercom I got a resource that is 0 bytes and 0 ms. That is incredible fast!
Maybe we could just skip resources that are 0 bytes / 0 ms from the fastest resource list, because it seems in relevant?

Best
Peter

Testing with www.facebook.com gives data:font/opentype;base64,T1RUTwAJAIAAAwAQQ0.... as fastest resources.

Needs to break the loop and any hanging phantomjs process when a page exception put phantom into a hanging state:
https://github.com/wesleyhales/speedgun/blob/master/server/src/main/java/com/fluxui/jms/PerfQueueManager.java#L296

Should add a way for users to specify the file output, and specify the url to post to.

Currently building a new container every time

Currently, any speedgun server instance has the ability to send a beacon signal to the centralized speedgun.io server hosted on digital ocean. This is an opt-in feature and it should be exposed and documented.

To initialize the beacon on a local running instance:
http://localhost:8081/rest/beacon/init

This will send a report to speedgun.io:

• cpu/memory usage
• overall availabiltiy info

Speedgun.io then stores this data as an available node:
http://speedgun.io/rest/beacon/getlist <-- This should be exposed on the UI as a dropdown list. It will allow users to run speedgun in a distributed fashion across many different instances and geographic locations.

Everything looks ok on the start page with the text and the design, however something is breaking it because you can scroll to the right and get a big white area.

I'm using an Iphone 5 and Safari with iOS 8.1.1.

Checkout the monster image :)

Best
Peter

Currently the description of what each API call does is embedded in the speedgun script. This should be abstracted out to the UI. It currently weighs down speedgun reports and is not flexible.
Example:
https://github.com/wesleyhales/speedgun/blob/master/core/speedgun.js#L79

It should be an executable shortcut (see CasperJS). An obvious solution is to make it npm install-able.

I tried to run the example that is on the README file and I got this error:

$ phantomjs --config=core/pconfig.json core/speedgun.js http://www.cnn.com -o csv --screenshot
TypeError: undefined is not an object (evaluating 'config.userAgentAliases[config.userAgent]')

  phantomjs://code/speedgun.js:437 in load
  phantomjs://code/speedgun.js:73 in run

It would be cool if the timings would be sorted per domain to be able to spot slow domains for third party
content.

So we would take the time per domain, use the median value and just list it something like this:
Slowest domains:

• gfx.omni-cdn.se 1203 ms
• ads.easy-ads.com 1171 ms
• s1.adform.net 980 ms

Best
Peter

It works, but it's time to retire and rewrite the entire backend for speedgun.io in something else. Ideas?

This can be cleaned up to handle virtualbox errors better... any new install has to run ./fix-virutalbox.sh

It seems like Speedgun isn't in the npm registry anymore.

npm install -g speedgun
npm ERR! Darwin 14.3.0
npm ERR! argv "node" "/usr/local/bin/npm" "install" "-g" "speedgun"
npm ERR! node v0.12.5
npm ERR! npm  v2.11.2
npm ERR! code E404

npm ERR! 404 Not Found: speedgun
npm ERR! 404
npm ERR! 404 'speedgun' is not in the npm registry.
npm ERR! 404 You should bug the author to publish it (or use the name yourself!)
npm ERR! 404
npm ERR! 404 Note that you can also install from a
npm ERR! 404 tarball, folder, http url, or git url.