Book Image

The Road to Azure Cost Governance

By : Paola E. Annis, Giuliano Caglio
Book Image

The Road to Azure Cost Governance

By: Paola E. Annis, Giuliano Caglio

Overview of this book

Cloud teams and ICT cost controllers working with Azure will be able to put their knowledge to work with this practical guide, introducing a process model for structured cost governance. The Road to Azure Cost Governance is a must-read if you find yourself facing the harsh reality of monthly cloud costs gradually getting out of control. Starting with how resources are created and managed, everything you need to know in order to track, display, optimize, rightsize, and clean up cloud resources will be tackled with a workflow approach that will leave the choice of operation to you (be it the Azure CLI, automation, logic apps, or even custom code). Using real-world datasets, you'll learn everything from basic cost management to modeling your cloud spend across your technical resources in a sustainable way. The book will also show you how to create a recursive optimization process that will give you full control of spending and savings, while helping you reserve budget for future cloud projects and innovation. By the end of this Azure book, you'll have a clear understanding and control of your cloud spend along with knowledge of a number of cost-saving techniques used by companies around the world, application optimization patterns, and the carbon impact of your cloud infrastructure.
Table of Contents (14 chapters)
Section 1: Cloud Cost Management
Section 2: Cloud Cost Savings
Section 3: Cost- and Carbon-Aware Cloud Architectures

Paying for consumption

One of the main obstacles to great Azure cost governance—and, to be honest, to any cloud cost governance—is the old IT misconception that a VM is something known, beloved, and cherished, much like a pet; while in the cloud, a VM is just a service, a commodity, and should be switched off and on according to its real usage.

According to the post at, servers that are given a friendly name and are considered indispensable and unique objects with critical support are what we call pets; they are manually operated and their owners have a clear affection for every single one of them. If they fail, everyone knows they are down, and the IT folks must immediately work on restoring them until they are back and shiny.

On the other side, arrays and systems built via automation, such as clusters, racks, and anything that is load-balanced, are what we call cattle; they can be switched...