Posted by Chad.Cumba on February 25, 2010 at 8:59pm
So I've asked around some of the Drupal boards a bit, and I can't find anyone that has a solid handle on a easy to use drupal webserver health reporting module. Currently I switch back and forth between Nagios and Cacti depending on client needs, but what I'd love to see is a drupal module that can just check simple stuff like cpu load, ram, swap, caching, and webtraffic load. There's a nagios drupal module, but that requires setting up Nagios and that takes more setup time than a lot of my clients want to bother with. Anybody got any ideas? Or should I just get off my rump and make one?
Comments
Monitoring web servers from
Monitoring web servers from Drupal is a bad idea. That's why there aren't any modules.
why bad?
Monitoring the webserver it's on from a drupal instance would be a bad idea. Monitoring a bunch of drupal sites on different webserver instances from one central location is an obvious idea.
It seem clear that there
It seem clear that there could not be a robust monitor (as drupal would be one of the first thinks to go when things get bad), and not a good solution for site that need excessive uptime. But I do not see why a module that fires at cron and runs a set of test would not be useful. especially if it provide a way to store such reports.
It would be nice to know that your server is slow on Tuesdays at 10. Even more so if you are on shared hosting and do not have control over the server or access to install a more complete solution.
While the tool would not guarantee notification of issue, it could still supply useful info, when one can not or is not willing to install the big boys
More what I was getting at
More what I was getting at yeah. Basically my issue is the users that don't see the value in paying for all the configuration of Nagios or Cacti or any of the real robust monitoring solutions. However, I'd still like some kind of easy ping when things start to go sideways. Heck, for a lot of low traffic user sites, a lack of that ping would signal that something is wrong.
Certainly it wouldn't be nearly as accurate, but if it was just prepping and sending an email that said "Hey the site is still alive and the logs aren't crazy, pages seem to be loading at ok speeds" that's better than nothing. Things like ram, cpu load etc would almost certainly need to come from outside services. Ie, if you're attempting to check them from PHP you'll have permissions issues, reliabilty, completely unscientific data sample sizes, just to name a few issues off the top of my head.
http://empoweredby.net
simple drupal heartbeat/check-in
It's a good idea.. If you have Drupal itself running these things (cron hook should call it, but drupal should do the legwork), you are engaging all the components in one. For the extreme opposite example of a bad idea, you wouldn't want some external wget which just checks for an http OK status, because your DNS might have been hijacked, and your domain could be serving up some hitfarm ads, but the heartbeat monitor will never know the difference.
We actually created a
We actually created a solution that provide this insights. You can find more information on www.drupalmonitor.com.
--
netnode.ch - Drupal Entwicklung
There definately are modules out there
How about:
http://drupal.org/project/droptor
Or as mentioned before http://drupal.org/project/nagios in combo with http://drupal.org/project/prod_check, although this combo does not provide performance monitoring... yet... I'm working on integrating the revived http://drupal.org/project/performance into prod_check as well.
And the integration is there.
And the integration is there. Should add an alpha version, but you can use prod_check from git + latest release of performance to have a combination that will allow performance monitoring without the need of Nagios.
Edit: Performance monitoring available for a while now in latest prod_check and performance modules.