Is apache running?

[squinky86]

One of my complaints for WHB is that the uptime system monitors whether a server replies to a ping or not. To fix this, I have been working on some php scripts to look at apache's uptime instead of the network's uptime.

I uploaded a small portion of it to my site: http://www.squinkyhosting.com/uptime

This system checks how quickly WHB's servers take to start responding to a simple HTTP GET request on a plain html file. Note that this does not measure php execution times or whether the SQL server is overloaded. It should bail out pretty quickly if the server is dead, and if apache is overloaded, you can see the lag time go up to 20 seconds (at which time, a timeout is recorded).

I tried to make the scripts as efficient as possible, but since it's using the GD library, it may get a little resource intensive. As such, please don't refresh my page(s) too often.

[squinky86]

Applause to WHB: from my intermittent, random monitoring, apache has been stable and snappy on all servers today . It hasn't taken more than a quarter of a second to respond to a static inquiry for any server.

[Tony]

Interesting, in an academic sense I guess. But, being on shared hosting, I'm much more interested in the real world response time of my website, and usually that has nothing at all to do with Apache.

I mean, Apache can be up and fine for 6 months, but one bad hop or one runaway script on the server can bring things to a crawl. I don't need a script to tell me things are crawling.

Now, if you can write a script that will give the cause of slowdowns, get back to me!

[Matt R.]

Monitoring Apache isn't necessarily accurate either, though. With a clean Apache restart, there is literally milliseconds that the site won't respond (most browsers won't even notice) but it would be shown as an Apache restart, etc.

[squinky86]

I modified the script to make sure it was getting "200 OK" responses from apache now.
You guys are right, it's still not accurate, but it is better than pinging. I think I already qualified this by saying that php execution time or SQL server usage are not measured. Unless I have an account on each server, it's going to be very difficult to measure what's overloaded.