HTTP Parse Check – Monitor Anything!

Most of the checks that we have released for NodePing to date have monitored specific Internet-based services. These are very useful, and they are the bread and butter of traditional web site and server monitoring services. Today we’re really excited to announce the release of our “Monitor Anything” check. You can use it to monitor… well, just about about anything.

The new check is the HTTP Parse check. It connects to a remote HTTP service and parses the response for specific fields and values. The values are then stored in the check results for reports. If any of the values are outside of your configured ranges, it triggers a notification via email, SMS, twitter, voice call, or webhook.

The HTTP Parse check can look for named fields in text or a JSON response. For example, if the check is configured to look for a field named “llamas” and the response contains llamas: 34 then the system will store 34 as the value, which will be displayed on reports and charts.

Similarly, the response could be in JSON format. For example, the JSON could be something like this: {"animals": {llamas": 34}} Configuring the check to look for animals.llamas will store 34 in the check results. The check can handle multiple fields, as long as they are all accessible from one request to a URL, and it can pull values from the middle of a larger response.

What could this check be used for? Anything that is accessible by an HTTP request that returns a field with a numeric value. Since it is parsing the response for the field name and value, it can be embedded in other data. This means that any JSON API or data service on the Internet is potentially a target for this.

Our use of this check at NodePing is to monitor server health. As you can imagine, running a service like NodePing means we have a number of servers all over the place. We use this check to monitor server load, free memory and disk space on a number of our servers.

There are many ways to generate the system information from a host so it can be used for this check. Some of the ones we’ve tried include Linfo and phpSysInfo. We wanted a similar tool written in Node.js, so we wrote npsystats, which is what we now use for our systems. We’ll have more to say about npsystats in the near future. There are a number of other similar tools out there, such as Ohai in Ruby. The HTTP Parse check is capable of working well with most of them.

For example, we have a check on a web server that monitors server load. npsystats returns the following json in response to the check’s HTTP GET request:
{"load":{"1min":1.86376953125,"5min":0.9501953125,"10min":0.64404296875}}
The check is set to watch three fields: load.1min, load.5min, load.10min, with a range of 0 to 8 (it could be different for each field, but it doesn’t have to be). So if the load on this server goes over 8, I will get an email. I also can access a graph showing the 1, 5, and 10 minute load for this server.

More information can be found in our documentation for the HTTP Parse Check. We think this makes NodePing an even more awesome piece of your system monitoring and automation toolkit, particularly when tied to judicious use of notifications and webhooks. If you don’t have an account on NodePing yet, try it out with our 15 day free trial.