Cloud FinOps, 2nd Edition

Chapter 22. Metric-Driven Cost Optimization

Metric-driven cost optimization (MDCO) is the driver that measures potential optimizations and then uses goals and target lines to trigger an operate process to action them. When you are buying commitments, you’re in the operate phase. With MDCO, you use the metric threshold to determine precisely when you take those actions.

MDCO could also be described as “the lazy person’s approach to cloud cost management.” The primary rule of MDCO is: don’t do anything. That is, don’t do anything until you have a metric that measures the impact of your actions. If you make optimizations but don’t have metrics to tell you whether they’ve been positive or negative on your spending, you’re not doing MDCO.

By the end of this chapter you’ll know how to use metrics to correctly drive your cost optimizations.

Core Principles

There are a few core principles that define the MDCO practice:

Automated measurement

Computers perform the measuring, not humans.

Targets

Metrics without targets are just pretty graphs.

Achievable goals

You need a proper understanding of data to determine realistic outcomes.

Data driven

Actions aren’t driving your data: the data is driving you to take actions.

We’ll explain each of these principles throughout this chapter.

Automated Measurement

Loading billing data, generating reports, and generating recommendations for optimizations are all tasks that should be done by automation, or via a FinOps platform. When massive amounts ...