At our company we have an internal https site that our staff log into each day. This site is load balanced (using Microsoft NLB) over two web servers and each web server has a dedicated application server it talks to. I was wanting to know if it would be possible to use PRTG to not just monitor IIS on the web server to see if it is running but instead be able to monitor the site itself to tell if the site loads properly or if tell if the site instead is showing an error in case there is an issue with IIS or in the application itself. Preferably I would also like it to login to the site to test if validation works too but that may be asking for the moon.
Then also (if possible) have PRTG send a command to our Microsoft NLB server to remove a specific host if testing the site comes back with negative results and then also add the site back into NLB if it test again and gets a positive result.
I think it can be done but just not sure where to start as there are many HTTP sensors and I really not sure which one I need.
What kind of failure are you looking for? Is there a specific error message that will be returned when the site is down or will it be a normal 404 / 500 message? In the case of a normal HTTP failure message, the normal HTTP Sensor would work for you since it would go into an error state when the site isn't available.
If you want to check the login, it depends on the type of login that you use but with the HTTP Advanced sensor to post data to the website to have the sensor log in. Once the sensor has logged in, you can have it check for specific content to make sure it worked properly.
Removing a host might also be possible if you set up a notification to do so. If testing the login came back false and the sensor goes into an error state, you could set up a notification to run an EXE or script to take out one of the hosts or if API calls are possible, you could do that through executing an HTTP action. You can see the details of the various notification methods here.
Jan, 2015 - Permalink