Improving Performance with New Relic APM and Insights

In any kind of tech development, knowledge is power. When working on an ecommerce site, knowledge is essential.

The more information you have about your application, both during development and when it goes live, the more value you will be able to provide your client and in turn to the client’s customers. When in a development environment, it’s easy to provide yourself with a breadcrumb trail back to an issue, but when your code moves into a staging environment, the information you provide can end up being a lot less useful. At one point, this was as useful as it got for us:

With no way to know what this “something” was, and after a few awkward problems where we very quickly reached a dead end, we made the decision to introduce New Relic APM into our workflow.

New Relic APM helps you monitor your application or website all the way down to the code level. We have been using this in conjunction with New Relic Insights, their Analytics platform.

With New Relic we have been able to track VPN downtime, monitor response times and get stack traces even when working in a Production environment. So the above vague message, becomes this:

This monitoring enables you to increase confidence in your product in a way that isn’t possible with simple manual or even automated testing.

In addition to the APM, we’ve also been working with New Relic insights. It behaves similarly to Google Analytics. However, its close ties to APM’s tracking and monitoring, means that the data is only limited by the hooks you create and the queries you can write in NRQL (New Relic’s flavoured SQL language). It feels far meatier than GA, and you can also more easily track back end issues like time outs, translating it into graphical form with ease (if you’re into that sort of thing).

Being a new product, it is not without its pitfalls. In particular, NRQL can feel quite limited in its reach. A good example of this is the much publicised addition of maths to NRQL. That a query language didn’t include maths in the first place felt a bit like an oversight. However, this has been remedied, and they have also introduced funnels and cohorts which should add a lot to the amount you can do using Insights.

As a company Red Badger has always valued fast, continuous development. While traditional BDD test processes have increasingly slowed us down, by improving our instrumentation integration we hope to be able to improve our speed and quality overall.