optimization

We will be doing a four-part blog series about looking at the sources of metrics for optimization. To kick off the series, I will teach you what optimization actually means, what you are trying to optimize, and if you are focused on business goals or technology goals when performing this optimization. 

In our next blog, we will look at APM derived metrics and metric domains you can leverage, as they are the two primary sources people rely on for optimization today. In our following blogs, I will cover why you might want to use one source over another and how to get to a consistent implementation for those resources. 

First Things First, Optimization Can Mean Many Different Things To Many Different People. 

To make sure everyone is on the same page regarding what optimization you are talking about, you need to ask, “is this a business goal?” or “is this a technology goal?” when talking about optimization. Let’s say someone wants to maximize their customer satisfaction. This type of optimization is considered a business goal. On the technology side, a goal with a similar intent would be how to minimize latency per transaction. These goals are known as, and can even be programmerly described as, a service level objective. 

Service Level Objectives 

Actual service level objectives would be something like wanting to ensure your customer’s satisfaction score stays at or above 4.5. You need to have a way of taking whatever that number 4.5 means to you and your organization and deriving that from a set of metrics that you can collect from your application space. It doesn’t help to try to optimize your latency to address a customer satisfaction goal unless you have a way to correlate those two results. This process requires some form of automated metrics gathering so you can tie it into your optimization system, and through that correlation, optimize your service. 

Let’s say you are getting constant customer satisfaction information because your customer satisfaction information is not only derived by customer feedback, but also potentially because you are deriving it from application latency. You can say yes to optimizing latency for your service level objective because you know latency correlates to customer satisfaction. This is typically done in the manual process of customer feedback and then correlating that latency; the goal is to have the two metrics come together in some kind of programmable connection. 

Now that you understand what optimization means and how to determine the goals your enterprise is focused on, tune back into our next blog to learn why you should use one source of metrics over another and how to get a consistent implementation for those resources.