HEAD vs GET Request...

Developer
Apr 17, 2009 at 4:39 PM
Edited Apr 17, 2009 at 4:39 PM
I noticed the application is performing a GET command.  For the purposes of checking the availability of a URL it would be more efficient and less load on the destination servers if it used a HEAD request instead of a GET request.  In the case where the user has specified to gather the html then a GET would be appropriate but for general availability I believe a HEAD request would be better.


Developer
Apr 21, 2009 at 6:30 PM
I did some testing with modifying the request to use a HEAD request and found a considerable difference in the impact to the webserver.  I wrote up an article that summarized my testing if anyone is interested in the results.
Coordinator
Apr 21, 2009 at 6:53 PM
Hi

Very cool article and nice observation! :)

Did you ONLY changed the lines you write in your article or did you do more? If not, I will implement it and make a new release of WebMonitor with this included.

Only problem for me is that the latest source actually is a "work in process" and I doing some "proto typing" on implementing MEF as a way to support plugins, so dev could write there own "rules" for the application. Sadly something else came up and I didn't got it finished.

Don't know if I got time now, but I will sure put in your fix for next release.

Really happy to hear your feedback, always good to hear people using and improving your work, good job!

//Qbus
Developer
Apr 21, 2009 at 7:12 PM
Edited Apr 21, 2009 at 7:18 PM
Glad you liked the article.   I think you have a pretty cool product.

I only changed the lines in my article and everything seemed to work fine...  What Source Code version is the one associated with the release version on the download page?

Let me know if you need any additional help as I like this sort of thing...

Chris
Coordinator
Apr 21, 2009 at 7:21 PM
I just posted a post about your article onmy blog : http://laumania.net/post/WebMonitor-improvements-article.aspx

Thanks for the feedback, really nice to hear. As you can see it's long time since I created this project and actually it lived long before it got on Codeplex, but yeah it's a pretty neat tool and I still use it my self :)
There could be MANY improvements, as you just proved, both technical ones and UI ones.

If you would like to improvement it even more, I can make you a part of the project here at codeplex?

I will implement your fix as I wrote, but sitting at a slow laptop now, so it have to wait alittle :(

//Qbus
Coordinator
Apr 21, 2009 at 7:25 PM
Hmm, sadly I can't find or recall which changeset is the one used for the release... :(
It's not that the lastest source code doesnt work I guess, there just could be bugs...but at the same time, there could be fixes too :P
Coordinator
Apr 22, 2009 at 11:55 AM
One of my colleagues jsut pointed out, that this "HEAD" method is good in most cases. But there could be a situation where the web server, for some reason, doesn't handles "HEAD" request, but only "GET" and "POST". Therefore there should be an option to force GET. I know you can do that by enable HTML collection, but that's not logical, right?

Just wanted to point it out.
Developer
Apr 22, 2009 at 4:29 PM
Edited Apr 22, 2009 at 4:46 PM
Interesting point...  The HTTP RFC includes the HEAD request so this requirement would only be for those web servers which don't follow the RFC guidelines.  In this case I would think it would be a good idea to document the configuration of enabling the HTML collection as a way to handle that requirement though...