Join GitHub today
GitHub is home to over 36 million developers working together to host and review code, manage projects, and build software together.Sign up
Record webserver and platform #223
There is interest (https://phabricator.wikimedia.org/T86559#976694) in finding out how many windows users there are of MediaWiki. It would be good to use the response headers to gather this sort of information. Even if we only recorded IIS/Apache/nginx, that would help.
Hmm... for some reason it was/is not possible to fetch the data for the webserver. I think to remember that @thingles and I e-mailed about this in early 2013. No sure what the exact issue was. For a start the field for it is present in the "Website" template from the very start of its existence.
I don't understand why it isn't possible, but maybe there is something about the way the bots are written that I don't understand.
For instance, I see that a request to grab my avatar here results in a server response with the included header line:
For this, "Apache" would be appropriate. Maybe an additional field that says "Via: 1.1 Varnish".
A request on wikiapiary itself results in the following header lines:
For this, "nginx/1.7.1" would be appropriate.
A request to a known IIS server gives:
For this, "Microsoft-IIS/7.5" would be appropriate. Maybe an additional field that says "Via: NS-CACHE-9.3: 1".