Re: [Pharo-dev] image traffic

2014-02-25 Thread Esteban A. Maringolo
Great stats.

AFAIK, approximately two thirds of the downloads are performed by curl UA.
Which means many people, including me, must be using the convenient (but
unsafe) http://get.pharo.org shell script.

Considering a CI environment or scripted automated build, many people is
accountable for more than one download per day (I'm for two the last days).

Anyhow, it is good to have some raw metrics to analyze on.

Regards!



Esteban A. Maringolo


2014-02-25 11:07 GMT-03:00 Camillo Bruni camillobr...@gmail.com:

 We finally have our stats for files.pharo.org online again:

 http://files.pharo.org/teststats/2014/02/awstats.awstats.en.html

 We're approaching around 300GB downloads per month, not bad I'd say :)



Re: [Pharo-dev] image traffic

2014-02-25 Thread Norbert Hartl
If you look at the hosts list than you can see that 75% of the overall traffic 
is generated by cloud-gw.inria.fr.  The CI jobs download the vm with curl.

What brings me to the next point. The „Hosts“ section in those states might 
break a data protection law. I don’t know how they are in france but in germany 
this would be illegal. We are not allowed to store IP addresses for a longer 
period. And making them public is really asking for trouble. 

Why? Because now I know that Sebastian Sastre might be responsible for 2,5% of 
the overall traffic on that site. I find that funny but most people don’t. 
Especially I’m asking myself why his server is in germany close to my own 
server maybe in the same room…. :) 

Anyway, an advize would be to check the law situation or just to switch off the 
Hosts section in the stats.

Norbert

Am 25.02.2014 um 16:54 schrieb Esteban A. Maringolo emaring...@gmail.com:

 Great stats.
 
 AFAIK, approximately two thirds of the downloads are performed by curl UA. 
 Which means many people, including me, must be using the convenient (but 
 unsafe) http://get.pharo.org shell script.
 
 Considering a CI environment or scripted automated build, many people is 
 accountable for more than one download per day (I'm for two the last days).
 
 Anyhow, it is good to have some raw metrics to analyze on.
 
 Regards!
 
 
 
 Esteban A. Maringolo
 
 
 2014-02-25 11:07 GMT-03:00 Camillo Bruni camillobr...@gmail.com:
 We finally have our stats for files.pharo.org online again:
 
 http://files.pharo.org/teststats/2014/02/awstats.awstats.en.html
 
 We're approaching around 300GB downloads per month, not bad I'd say :)
 



Re: [Pharo-dev] image traffic

2014-02-25 Thread Christophe Demarey

Le 25 févr. 2014 à 17:37, Norbert Hartl a écrit :

 If you look at the hosts list than you can see that 75% of the overall 
 traffic is generated by cloud-gw.inria.fr.  The CI jobs download the vm with 
 curl.
 
 What brings me to the next point. The „Hosts“ section in those states might 
 break a data protection law. I don’t know how they are in france but in 
 germany this would be illegal. We are not allowed to store IP addresses for a 
 longer period. And making them public is really asking for trouble. 

I think we have some equivalent laws.
For another project, we had to anonymize awstats results.
Maybe raw statistics should be kept private and only use numbers for marketing.

Christophe.



smime.p7s
Description: S/MIME cryptographic signature


Re: [Pharo-dev] image traffic

2014-02-25 Thread Pharo4Stef

On 25 Feb 2014, at 17:48, Christophe Demarey christophe.dema...@inria.fr 
wrote:

 
 Le 25 févr. 2014 à 17:37, Norbert Hartl a écrit :
 
 If you look at the hosts list than you can see that 75% of the overall 
 traffic is generated by cloud-gw.inria.fr.  The CI jobs download the vm with 
 curl.
 
 What brings me to the next point. The „Hosts“ section in those states might 
 break a data protection law. I don’t know how they are in france but in 
 germany this would be illegal. We are not allowed to store IP addresses for 
 a longer period. And making them public is really asking for trouble. 
 
 I think we have some equivalent laws.
 For another project, we had to anonymize awstats results.
 Maybe raw statistics should be kept private and only use numbers for 
 marketing.

+1

 
 Christophe.