Teamness blog: Monitoring a web application

So we got a reliable web hosting service for Teamness and we came up with a security policy to keep the data safe.

All good, but we also needed a monitoring solution, to be informed when something is wrong with the server.

We settled to use Pingdom, a Swedish service built for this purpose. Big guys like Alexa, iStockPhoto, Feedburner and Skype are using Pingdom, but it's also an affordable solution for small startups like us.

This post is not a Pingdom review, but more about the way we're using their services to know that Teamness is up and running.

Our Pingdom account

We set up 3 monitoring options for Teamness, for every web application in Teamness:

the public website
the private website
the administrative web application that we use internally

For each of the above, there is a check done every minute. If 2 consecutive checks fail, each of us gets an email and an SMS. We're not all the time in front of the computer, but we carry our mobile phones everywhere, so we know in an instant if something is wrong.

There is also the option to pause a checking routine, so the notifications are off while we're doing maintenance work, like uploading a new build.

What do we check for?

We built a special tiny module in Teamness, which performs a couple of sanity checks to be sure the application is running safe. This module is called from a separate page, designed specifically for this purpose, which is further requested in the Pingdom checking routine.

This way we don't only know that the server machine is up and that the web server is running, but also that the inner workings of the application are sane.

Current status pages

You may expose the status of your web application through a public link. We used TinyUrl to generate these custom links for the two websites in Teamness:

They are very useful when our users ask for the uptime. A link to a nice page with the historic of the availability is more valuable than stories and explanations.

The link for the private website, which is the website with your data, is published on the Teamness main page.

Pingdom goodies

They have a set of tools for checking a webpage, by mimicking the way it is loaded in a web browser and also for doing traceroutes or pings.

Tests may be saved for future reference, like this one for The New York Times.

Mr Uptime is a Firefox extension that will let you know when a website you checked and wasn't available is again back in business.

Royal Pingdom is the name they chose for their blog. I like the Swedish humor. And I like the blog, is one of the corporate blogs that issue quality posts and not only blabber about some new offers.