check_webpage.sh

Project Notes

DESCRIPTION:
The principal object of check_webpage.sh is to give stats about the time a complete webpage takes to load. You can monitor your own webpages or you can take some general stats of your navigation infraestructure. By default the script always gives an OK state because it has statistical purposes, but is prepared to give you warning or critical states depending on the values you want for elapsed time or percentaje of erroneous downloaded files
The script basically use wget to obtain an entire webpage with her own dependences (css, images, etc). I prefer wget instead curl because it has more recursion support to parse and obtain dependences. The problem is that wget download the index.html and parse it to obtain the rest of dependences, so you need to have this file in somewhere to download de rest. If you have to exec various checks at same time, all the work can be used in I/O opperations and you are going to take bad results in your stats. If you try to redirect all to /dev/null you can't parse the index.html and the download end. To solve this problem I use a TMPFS File System of 6 Mb that is unmounted when the script ends or if we receive some
TERM signal

Use -h or --help to see the options it can use.

If you are getting some trouble, use the -D (Debug) option to see what the script are doing. The Debug output is sent to STDERR

IMPORTANT: Read the comments in the script to see how it works and the necessary configuration dependences of your system ;)