How to monitor Web server downtime?

I have a new Linode with CentOS 5.x and Lighttpd web server.

Do you have any tip or trick about how to monitor web server uptime or downtime? I want to make sure the web server is up all the time or notify me in case there is any problem.

What do you suggest?

Thank you.

Julian

15 Replies

Nagios is by far the best network monitoring tool in my opinion.

It kind of relies on the idea that you can set up a Nagios server somewhere other than the server you want to monitor (not mandatory but it makes sense, if the server goes down, a local Nagios will not be able to tell you about it!), then you set up periodic checks of your web server, and trigger alerts (email, sms even) when something goes wrong.

Installing Nagios on the actual server you want to monitor can also be useful as it will by default monitor the local CPU, disk space and so on.. you can set this up from an external nagios server but it involves a bit more configuration at both ends.

If you want something a little bit lighter and quicker, you could just write a script and run it on a cron to check your HTTP is responding. For instance I've written perl scripts that use curl to load a website every 5 minutes and check to make sure a certain keyword appears on the page, and if not, e-mail me.

something simple, free and quick is to use one of the free website monitoring services. i use site24x7.com and also siteuptime.com. both monitor ay specified time intervals (free intervals start at every 15 minutes) and alert via email if site is down. i think you can pay to get quicker checks and also sms alerts.

You might look at monit, which would also be able to restart services automatically.

I've started using Zabbix, but things like Zabbix and Nagios are designed for monitoring lots of hosts rather than just a single node.

@phvt:

I've started using Zabbix, but things like Zabbix and Nagios are designed for monitoring lots of hosts rather than just a single node.

True I didn't consider that fact.

I think pingability.com also offer these sorts of services though there are conditions on its 'free' use.. totally offtopic but their DNS check is excellent :)

I'm a big fan of Nagios but for single / small applications I've used http://basicstate.com

Nice setup and easy to use.

Good luck!

Thank you for sharing the information.

I think I will try any of the free external services until I need a professional service.

For a larger network, I setup Big Sister to monitor everything. It can monitor more than just your web server.

Might be overkill for just one machine, but if your network grows, something to look at.

For monitoring I highly recommend a hosted solution, so it is completely independent of you and your network.

There are only few good free ones. I am currently with AlertFox on the Free2 plan and it works great - especially the free transaction monitoring. I use it to monitor our user login and image upload functions. Because on a server much more can go wrong than just a simple downtime ;)

Ooops, double posted => deleted this entry (see above for real post) :oops:

Hi, Iv just built a server monitoring web application. There are no free plans at the moment but it quite reasonably priced compared to a lot of the services I have seen. It checks your sites from 2 different locations and checks to see if the html contains a phrase or keyword that you can setup.

One of the monitoring nodes is hosted at Linode in their Newark data centre.

The site is called Runs Like Clockwork and is at http://runslikeclockwork.com

We have a presence in all four Linode datacenters and also offer a similar service at http://www.SiteSteady.com/

We're currently working to wrap-up beta testing. jmagnone, send me a PM (private message) and I'll set you up with a free account.

I've been using the free service of site24x7.com for a few months now. What I like about them is their response time graphs. For example, I could immediately see the difference in response times after I migrated a website to Linode. (If you're curious, average response time dropped by about 100 ms from my self-hosted server here in Europe, but there are infrequent peaks where response time can be as slow as 3–7 secs which I didn't have before. This is a static site served by nginx running on a Linode 360 in the Dallas DC.) Also, the site24x7 "site is down/up" email notifications seem to work flawlessly (no down time on Linode so far).

There's something about the organization and overall feel of site24x7.com that doesn't excite me, though, which is why I continue to keep my eye open for other solutions. What I'd really like is a service that uses the push notifications of iPhone 3.0.

By the way, just-ping.com is a very cool ping service.

I just had my linode lock-up and take down all the sites on it. CPU spiked to 400% and remained there. What I would have liked to have happened is for the server to have automatically restarted itself. Is there tools available that would handle this? Lassie is enabled on the linode but that just seems to be for power outs.

Thanks.

Robin.

You likely ran out of memory, leading to the invocation of the dreaded OOM.

http://www.linode.com/wiki/index.php/RebootingonOOM

@robin:

I just had my linode lock-up and take down all the sites on it. CPU spiked to 400% and remained there.
I would seriously look at fine tuning your config files a bit first, before enabling the automatic panic/reboot on OOM. [Install munin or similar, prune Apache (or install Nginx) and Mysql, etc. etc.]

Just my 2c.

Reply

Please enter an answer
Tips:

You can mention users to notify them: @username

You can use Markdown to format your question. For more examples see the Markdown Cheatsheet.

> I’m a blockquote.

I’m a blockquote.

[I'm a link] (https://www.google.com)

I'm a link

**I am bold** I am bold

*I am italicized* I am italicized

Community Code of Conduct