Wikimedia content can be used out of Wikimedia servers through the MediaWiki API. For instance, we can see Wikimedia content in various web services and mobile apps.
Do we have any metrics about the API requests we are getting in Wikimedia servers?
Yesterday a developer asked me (for the sake of discussion) how would we feel if the traffic in Wikimedia servers would decrease 1% while the content served through our API would grow 10%. I found the question very interesting, and made me wonder what do we know about our API usage.
This hasn't been studied yet AFAIK, although Erik Zachte might have looked at API usage in the past. We could analyse the number of API requests from the sampled1:1000 webrequest stream. What type of metrics are you particularly interested in? D
On Sat, Jul 6, 2013 at 10:42 AM, Quim Gil qgil@wikimedia.org wrote:
Wikimedia content can be used out of Wikimedia servers through the MediaWiki API. For instance, we can see Wikimedia content in various web services and mobile apps.
Do we have any metrics about the API requests we are getting in Wikimedia servers?
Yesterday a developer asked me (for the sake of discussion) how would we feel if the traffic in Wikimedia servers would decrease 1% while the content served through our API would grow 10%. I found the question very interesting, and made me wonder what do we know about our API usage.
-- Quim Gil Technical Contributor Coordinator @ Wikimedia Foundation http://www.mediawiki.org/wiki/**User:Qgilhttp://www.mediawiki.org/wiki/User:Qgil
______________________________**_________________ Analytics mailing list Analytics@lists.wikimedia.org https://lists.wikimedia.org/**mailman/listinfo/analyticshttps://lists.wikimedia.org/mailman/listinfo/analytics
although Erik Zachte might have looked at API usage in the past
No, I have not looked into this.
Even when online requests are lower by 1% and API requests on the rise with 1% rather than 10% there might be an inherent multiplier at work.
I mean one API request could lead to many article views via apps.
Some smartphone/tablet apps provide offline access to our full content base. These probably all use the dumps.
But some apps provide content specific topics, these might use the API for navigation or even collect all raw content per page, rather than via rendered html.
Erik
From: analytics-bounces@lists.wikimedia.org [mailto:analytics-bounces@lists.wikimedia.org] On Behalf Of Diederik van Liere Sent: Saturday, July 06, 2013 5:27 PM To: A mailing list for the Analytics Team at WMF and everybody who has an interest in Wikipedia and analytics. Subject: Re: [Analytics] MediaWiki API metrics of Wikimedia sites?
This hasn't been studied yet AFAIK, although Erik Zachte might have looked at API usage in the past. We could analyse the number of API requests from the sampled1:1000 webrequest stream. What type of metrics are you particularly interested in?
D
On Sat, Jul 6, 2013 at 10:42 AM, Quim Gil qgil@wikimedia.org wrote:
Wikimedia content can be used out of Wikimedia servers through the MediaWiki API. For instance, we can see Wikimedia content in various web services and mobile apps.
Do we have any metrics about the API requests we are getting in Wikimedia servers?
Yesterday a developer asked me (for the sake of discussion) how would we feel if the traffic in Wikimedia servers would decrease 1% while the content served through our API would grow 10%. I found the question very interesting, and made me wonder what do we know about our API usage.
Especially with Wikidata we expect a rather bigger usage of the API, as the machine-readable data can be accessed through the API instead of the HTML page.
Even though we do not need to analyze it right away, I would love to know whether the raw access data is stored or aggregated anywhere?
2013/7/7 Erik Zachte ezachte@wikimedia.org
although Erik Zachte might have looked at API usage in the past****
No, I have not looked into this. ****
Even when online requests are lower by 1% and API requests on the rise with 1% rather than 10% there might be an inherent multiplier at work.****
I mean one API request could lead to many article views via apps. ****
Some smartphone/tablet apps provide offline access to our full content base. These probably all use the dumps.****
But some apps provide content specific topics, these might use the API for navigation or even collect all raw content per page, rather than via rendered html.****
Erik****
*From:* analytics-bounces@lists.wikimedia.org [mailto: analytics-bounces@lists.wikimedia.org] *On Behalf Of *Diederik van Liere *Sent:* Saturday, July 06, 2013 5:27 PM *To:* A mailing list for the Analytics Team at WMF and everybody who has an interest in Wikipedia and analytics. *Subject:* Re: [Analytics] MediaWiki API metrics of Wikimedia sites?****
This hasn't been studied yet AFAIK, although Erik Zachte might have looked at API usage in the past. We could analyse the number of API requests from the sampled1:1000 webrequest stream. What type of metrics are you particularly interested in?****
D****
On Sat, Jul 6, 2013 at 10:42 AM, Quim Gil qgil@wikimedia.org wrote:****
Wikimedia content can be used out of Wikimedia servers through the MediaWiki API. For instance, we can see Wikimedia content in various web services and mobile apps.
Do we have any metrics about the API requests we are getting in Wikimedia servers?
Yesterday a developer asked me (for the sake of discussion) how would we feel if the traffic in Wikimedia servers would decrease 1% while the content served through our API would grow 10%. I found the question very interesting, and made me wonder what do we know about our API usage.
-- Quim Gil Technical Contributor Coordinator @ Wikimedia Foundation http://www.mediawiki.org/wiki/User:Qgil
Analytics mailing list Analytics@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/analytics****
Analytics mailing list Analytics@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/analytics
On Sat, Jul 6, 2013 at 7:42 AM, Quim Gil qgil@wikimedia.org wrote:
Do we have any metrics about the API requests we are getting in Wikimedia servers?
This isn't all API requests, but we do know which edits to Wikimedia projects are made using the API, via EventLogging and Schema:PageContentSaveComplete. There's also information about the API application servers on ganglia, which might be relevant.