Re: Piwik tracking

Hey everybody

First, two good news.
I got the attention of somebody from the Piwik project and they would be happy to help us.
Piwik is running again on all profiles since this afternoon

I'm currently enabling Piwik and although my previous email (below) has answers, the following has still a need for answers.

Note that I am keeping the questions numbers the same as 'Re: Piwik tracking' thread [2] and [identifiers] are also kept.


Unfinished question

1. Why do we not track through all the domains…
I understand that we have a segmentation, but in some places, it is broken.

Piwik has a site configured to only track within 'docs.webplatform.org/wiki/'.

But is used in non shared files that are OUTSIDE of docs.webplatform.org/wiki/

Among others:
  - http://www.webplatform.org/
  - http://talk.webplatform.org/
  - http://www.webplatform.org/stewards/w3c/  (all stewards pages)
  - … probably others…

I would suggest we create a new site specific for the static pages as I did created new ones (see 'New sites' below).

New sites:
  Here are the two new websites that are now tracked within Piwik:
Project  (bug genie) [how:1]
Status [how:3]

Please tell me your opinion about others we might want.




New setup overview

New setup:
- 2 nodes configured as round robin under NGINX
- Reports are now archived hourly instead (reduces load) [5]
- DNS lookup are now against a local database that updates automatically
- Use of same database config as before
- Activated e-mail reports to be sent through sendgrid
- Archiving job sends message in Syslog



Tracking code:
- To use recommended integration instead of hardcoding [8], see [how:XX] for instructions
- Blog now has  WP-Piwik plugin (installed) [7]
- Docs to use Piwik_Integration [6]



Which Optimizations are done:
As described in [9], and [4]:
 - Auto archiving
 - Enable Geo location
 - NGINX, APC
 
But not (yet):
 - Data Limits
 - Secure Piwik
 - Use of MySQL Tuning primer
 - Use of GeoIP/NGINX instead of the flat file [10]


How I installed [how:XX]

[how:1] In BugGenie:
  - created a file in thebuggenie/themes/webplatform/tracking.js
  - called by: thebuggenie/themes/webplatform/theme.php  like the CSS file is
  - Visitor exclusion: TBD


[how:2]: In MediaWiki
  - created two files in code/docs/current/extensions/piwik/
  - added file: [13] "Piwik.hooks.php"
  - added file: https://raw.github.com/DaSchTour/piwik-mediawiki-extension/master/Piwik.php
  - Modified code/docs/Settings.php accordingly to [6]
  - Visitor exclusion: will not track authentified MediaWiki Administrator users [11], see list in [12]
 

[how:3]: Status (tumblr)
  - Login to Tumblr, in Settings, WebPlatform.org site, Search for Theme, click 'customize', find 'Edit HTML' button, added code.
  - Visitor exclusion: N/A


[how:3]: In WordPress
  - Installed [7] Wp-Piwik plugin
  - Visitor exclusion: will not track authentified WordPress with role Administrator [12]


[how:4]: In static
  - Edited in file: code/nonshared/webplatform/templates/container_close.html
  - Used on the pages:
    - http://talk.webplatform.org/
    - http://www.webplatform.org/errors/503.html
    - http://www.webplatform.org/errors/404.html
    - http://www.webplatform.org/stewards/adobe/ (and all other stewards)
    - http://www.webplatform.org/logo/
  - Visitor exclusion: N/A




Regards,

Renoir
~

Identifiers:
  [0]: http://project.webplatform.org/infrastructure/issues/INFR-48
  [1]: http://www.garron.me/en/blog/install-piwik-behind-varnish.html
  [2]: http://lists.w3.org/Archives/Public/public-webplatform/2013Sep/0052.html
  [3]: http://piwik.org/faq/new-to-piwik/#faq_134
  [4]: [ http://piwik.org/docs/optimize/how-to/ , http://piwik.org/docs/optimize/ ]
  [5]: http://piwik.org/docs/setup-auto-archiving/
  [6]: http://www.mediawiki.org/wiki/Extension:Piwik_Integration
  [7]: http://wordpress.org/plugins/wp-piwik/
  [8]: http://piwik.org/integrate/
  [9]: http://piwik.org/docs/installation-optimization/
  [10]: http://piwik.org/faq/how-to/#faq_166
  [11]: http://blog.webplatform.org/wp-admin/options-general.php?page=wp-piwik/wp-piwik.php&tab=tracking
  [12]: http://test.webplatform.org/wiki/Special:ListGroupRights (search 'protect', as [13] is checking)
  [13]: https://raw.github.com/DaSchTour/piwik-mediawiki-extension/master/Piwik.hooks.php


Original thread, [identifiers] are kept:

On 2013-09-16, at 12:08 PM, Renoir Boulanger <renoir@w3.org> wrote:
Subject: 
> (…)
> I'm working toward enabling again Piwik tracking throughout the site and I have a few questions in mind.
> 
> 1. Why do we not track throughout all the domains instead of specific per subdomain? Is it only to have a segmentation already made?
> 
> 2. Why do we use Fastly as Piwik frontend when it is technically useless since piwik big usage vector is to create stats (e.g. /piwik.php) and we cannot cache it through HTTP? [1]
> 
> 3. Who has the [superadmin] password, I do not want to override it?
> 
> In case you want to follow the progress, see this task [0] in the project management system.
> 
> I'll keep you posted.