We recently wrote a Nagios plugin for getting metrics from AWS Cloudwatch to LibreNMS. As the customer was ok with releasing it as open source here it is. It is written in Ruby - we like the language and we're also Puppet guys, so it felt like a natural choice for us.
The plugin allows getting as many metrics you want from Cloudwatch and setting alert thresholds (warning, critical) for each metric statistic (e.g. "average CPUUtilization percent") separately. In many cases it is not useful to configure alerts for a metric (e.g. "minimum CPUUtilization percent"), but you do want to graph it regardless - this plugins allows that.
The assumption behind the plugin is that you would monitor a single AWS resource (RDS DB instance, load balancer, etc) with it. This felt reasonable as you can have a separate service check for each resource.
For details check the project page for nagios-plugin-cloudwatch. We're sure the plugin can be improved a great deal, so please test-drive it and send us feedback via GitHub issues or better yet, pull requests.
