In this chapter, we deep dived into PM2 monitoring through Keymetrics. We learned how to put tight monitoring in place so that we are quickly informed about the failures in our application.

In the software development life cycle, the QA phase is, in my opinion, one of the most important one: no matter how good your code looks, if it does not work, it is useless. However, if I have to choose another phase where engineers should put more emphasis, it would be the deployment, and more specifically, the monitoring that is carried out after every deployment. If you receive error reports immediately, chances are that the reaction can be quick enough to avoid bigger problems such as corrupted data or customers complaining.

We also saw an example of active monitoring carried out by Netflix on their systems, which even though might be out of the reach of your company, it can spark good ideas and practices in order to monitor your software.

Keymetrics is just an example that fits the bill for Node.js as it is extremely well integrated with PM2, but there are also other good monitoring systems such as AppDynamics, or if you want to go for an in-house software, you could use Nagios. The key is being clear about what you need to monitor in the application, and then, find the best provider.

Another two good options for monitoring Node.js apps are StrongLoop and New Relic. They are both on the same line with Keymetrics, but they work better for large-scale systems, especially StrongLoop, which is oriented to applications written in Node.js and oriented to microservices.