In the first part of this blog series, we discussed the first three steps to cost control and visibility, and today we will focus on actively reducing costs. Here are three more steps you can take to drive the wise use of cloud resources in your organization, keep spending under control, and develop a plan to increase the value of your cloud investments. These are the next three steps:
This should be obvious, but it often isn't. Unused resources in the cloud cause - and this will come as no surprise - unnecessary costs. You should remove those resources. Classic examples from my daily work as a consultant are virtual machines from development and sandbox environments that have been shut down for months, disks that are not mounted, outdated snapshots, or unallocated public IP addresses.
Start with Azure Advisor, which will show you these resources and remove them - after consulting with the owner, of course.
Unlike an on-premises infrastructure, where it is virtually irrelevant whether a system is running 24/7 because the underlying server is always running, in the cloud this means that costs are incurred continuously.
Use simple mechanisms such as automatically shutting down a virtual machine at a specified time. This is especially useful for development and test systems. Show your staff how to prevent the shutdown if necessary, and how to restart the systems the next time they are used.
This simple measure can also result in significant cost savings. Imagine that a 730 hours/month (24/7) uptime can easily be reduced to 168 hours (8 hours on 21 working days per month). 76% less uptime and the systems are still available when you need them.
My customers who are moving from the on-premises world to the cloud for the first time tend to take certain "shortcuts" that they are accustomed to from the on-premises world. One of the most common is to go to the top of the shelf rather than the bottom when selecting machine sizes. In other words, the machines provided are often oversized. In some cases, a machine size is selected that is five to six times too large. This has a huge impact on costs.
Check the CPU and memory utilization of your machines in sequence, starting with the largest machines in your environment. If the average utilization is less than 50%, cut the machine size in half. Repeat this process until you find the optimal size.
I also recommend that you periodically switch from a series of machines that permanently provide the aforementioned CPU performance (e.g., Azure D and E series) to a B series. This series consumes CPU credits that are used under load and collects these credits when the machine is not in use. Ideal for systems that need to be available around the clock, but tend to be underutilized at night.
The steps described here are a promising start, but only the beginning of an ongoing FinOps journey. In the next phases, we should work together on strategies that maximize the value of your cloud investments.
Forward-thinking cloud management goes beyond cost reduction to recognize and strengthen the link between spending and business value. In the FinOps framework, this is referred to as "unit economics”. The goal is to identify the right metrics, effectively engage the team, and implement appropriate monitoring and forecasting.
These tasks go far beyond technical optimization, as we ultimately measure value creation in terms of profit margins achieved. Profit is the result of revenue minus unit costs. So far, we have only looked at costs. It is time for the next steps: Maximizing business value by leveraging the cloud.
We look forward to helping you get there.
Contact us now.
