Someone should create a service like that!

Some time ago Shawn and I were lamenting about what a pain network and service monitoring can be. There are some very good open source applications out there for doing this sort of thing. We’ve both used several versions of Nagios, and it works really well. If you need to run your own monitoring or write your own custom plugins (which we’ve done in the past), that’s a good option. If what you want is to monitor a bunch of services easily without having to put up and maintain another server just for that, a service that does it for you is more attractive.

There are a number of services out there that do pings, HTTP checks and a variety of other checks with notification. Some of them even start out at low cost or free, but if you have more than a handful of hosts and ports to monitor, they get pricey fast, or they don’t let you check very often, or they have some other catch that makes them just not do what you want. Some of them you need a special graduate degree from MIT to understand the pricing.

So we were trying to figure out how to get monitoring done reliably and cost effectively for a set of services we were responsible for at the time, and saying to each other “Someone should create a service that is easy, just does what you need it to do reliably, and doesn’t cost a lot.” Someone, as it turns out, was us.

More recently Shawn and I were once again chatting about the kinds of things geeks talk about, and one of those things was Node.js. I had been working on a few projects just as a proof of concept. It was clear that Node.js has some real strengths for writing scalable asynchronous services. In the course of the conversation, it occurred to us that we could create a service that would scale to many thousands of checks with very low incremental cost. If someone wanted to check thirty or fifty hosts every minute, the cost would be very similar to checking three sites every fifteen minutes. NodePing was born.

The name NodePing stuck with us, not because it uses Node.js (although it does, and we’re proud of that fact), but because “node” refers generically to something on the network. Of course, it’s much wider than that, and the most common checks don’t turn out to be pings. We think the name NodePing conveys “checks on things on the ‘net” well, even beyond pinging nodes. Our goal is to let you check what you want, when you want, for not much money.

As we wrap up our initial testing (with thanks to our beta testers for some great feedback) and move towards taking on customers in real quantity, I recognize that getting here has been quite a process from that first conversation about how lousy the options were for monitoring. I wish this service had been available when I was responsible for a range of web and email services years ago. It would have made life easier, for a great value at the price. We hope you see it that way too.

What do you want from a monitoring service? We are creating the service we wished we’d have had. What would you add? Is there something you’ve been frustrated about monitoring services, and just wish someone would fix already? Let us know!