Best practices in web server uptime monitoring

Trojan

I was wondering what you folks use or see as best practices when monitoring your sites, or clients sites, for uptime? Or more importantly, downtime

xaban

I've been using www.website-monitoring.com, which I have chosen after reading this article: "15 website monitoring services compared".

Cheers,

ocallagh

Pingdom are good - free SMS/Emails etc. I can tell it works as I had my first server go down on me last week.. got the texts from pingdom within about 20 seconds of it going down.

edit: Best practices:

3rd Party SMS/Email monitoring like above (check different ports, response times etc)
Cron which checks Apache+MySql are up and txtx/emails me with issues
Cron which checks a few DB timestamps to ensure the app is still ticking along

amen

don't forget though that if you are in a clustered environment you also need to monitor each of the nodes in the cluster, the response time of each node,

Also random sample log files and and have a history of response/access time on parts of the application/files.

Handy to see if something has changed (maybe a page displaying historical data and the query is no longer efficient that sort of thing)

If there are logged in users envolved also track response times vs number of simultaneous users

Trojan

Are you guys just checking HTTP is up, calling a static HTML page, calling a script that does other tests?

What about false positives? Some of the services I've seen are prone to false positives due to lack of multiple sites.