Ask anyone tasked with managing the availability of an application about their performance-based alerting, and many will likely tell you it’s tricky business.
On the one hand, if you set your alert threshold too low, you’ll get inundated with false alerts. And when that happens, you either waste time checking on each notification, or, after losing confidence in the system, start ignoring them altogether (which kind of defeats the purpose).
Yet on the flip side, if you set your alert threshold too high, you might miss a major drop in performance, which could result in significant end-user and business impact. So what’s the trick to setting your alerts just right?
At New Relic, we believe the answer lies in Apdex, an industry standard used to measure users’ satisfaction with the response time of an application or service. Unlike traditional metrics like average response time which can be skewed by a few very long responses, Apdex provides better insight into how satisfied users are by measuring response time against a set threshold.
We recently created a new white paper that examines Apdex and the alerting conundrum, which we encourage you to check out. There seems to be some misconceptions out there about Apdex, so if you want to get the real, technical explanation of how it works, the “Improve Alerting Accuracy” white paper is full of useful nuggets of information. You can also watch the video below from New Relic University:
For additional information on Apdex, take a look at: