Over the years, I have noticed an interesting and challenging problem when setting up transaction monitoring for various web applications. The root of the problem is the use of absolute paths in links, rather than relative paths.
For example, an absolute path would be "http://www.somesite.com/gotothispage.php" whereas the same URL in a relative path would be "/gotothispage.php".
When attempting to implement transaction monitoring that runs against individual servers in a load-balanced environment, the use of absolute paths makes this challenging. The reason is that the monitoring would point to the individual host, but the links, form submittals, etc. effectively redirect the monitoring to the full path of the application from the individual host.
To work around this, host files have to be used on the monitoring server, with the IP address for "www.somesite.com" being changed in between the transaction monitor runs against each individual server. This has it's own problems, in that the transaction monitors for various hosts may overlap and cause inaccurate results.
Developers: Please remember to us relative paths when writing code for web sites -- it will improve application monitoring, incident resolution and cause analysis a great deal.
Aug 11, 2015
Well... after about ten years of my site saying "revamping... new stuff to come," I finally got around to publishing a real site.
The site is mostly an online résumé at this time, and I will be adding some other information about me, my philosophy with regards to technologies and proceses, as well as some of my personal interests.
Generally, I'm a geek with interests in many technical areas, not just computers, networking, and systems. Geekiness can apply to many other things, such as taking pictures (I would not presume to call it photography, in my case), building things, gardening and growing yummy food, as well as cooking and eating yummy food.