Read the 2021 State of FinOps Report

Member Call: FinOps for Kubernetes Whitepaper Panel Discussion

Stacy Case
Stacy Case in FinOps Foundation
24th November 2020

To celebrate the launch of the FinOps Foundation’s recent Kubernetes whitepaper, we held a live panel discussion with 5 experts to answer member questions from the 295 participants who joined.

Unlike our typical member only calls, this session was open to the public as it was a collaboration with the Cloud Native Computing Foundation’s CloudNativeCon event. The full video is available above or in our member’s site.

We’re thankful to have everyone join this recent call, where a panel of Kubernetes and container experts and some of the contributors to our recent whitepaper got together to answer questions from the paper and on FinOps best practices for containers.

Foundation news and updates

Before we dig in, here are some foundation news and highlights.

About the panel and whitepaper

In case you missed it, we launched our first multi-member collaborative whitepaper on FinOps for Kubernetes. This was a huge collaboration by many foundation members and experts in the field.

The best part is that this is a living document– as cloud continues to change and become more complex, we’re going to update the insights here to keep everyone on top of FinOps best practices for Kubernetes and containers.

To ensure we cover all the nooks and crannies of FinOps for Kubernetes (and containers in general), we gathered the main authors and a few additional moderators together to answer member questions. Our panel included:

Notable themes and highlights

Container cost challenges are different than the rest of cloud infrastructure. Shared clusters can generate all kinds of cost data generated by different teams. These costs don’t line up “one to one” compared to other services like storage or VMs/compute. On top of this, shared container costs can add new wrinkles to how teams manage Kubernetes and containerization costs.

Container clusters break the mold of how IT normally looks at purchasing computing power. Since a container cluster can use a variety of number of resources to do work, it’s impossible to accurately track what certain teams are using by trying to allocate specific resources to them alone. You need to look at usage a different way since in a container context, infrastructure is shared and much more nuanced and dynamically used.

Trying to simplify container allocation is not the answer. You have to tackle it a much more scientific way– combining multiple sources of data to make sense of container costs, which from a FinOps standpoint means rethinking the way these container costs are labeled and accounted for. This means leveraging different ways to label container workloads and where they come from to get an idea of costs generated by teams. As one panelist points out, “it can be quite the journey for many teams.”

Strategy first: understand the workloads and resources that you actually take to Kubernetes, rather than waiting to make sense of a bill later. Some of our panelists discussed finding ways to scope and understand what those workloads will actually be to get a better sense of individual units of cost. You can then add nuance to this estimate by including shared costs and other infrastructure pieces that will go into those costs. The idea seemingly is to become more proactive about building a strategy to understand costs rather than just making sense of the “receipt like” data that typical cloud infrastructure generates. The goal is to find something to act as a reference or “to look for” when making decisions and sense of container costs.

… and many more insights. Please watch the full recording (embedded above, or you can watch it on YouTube here) for all the details and information. If you aren’t a member yet, please join us for free.

How to contribute to the whitepaper (and the FinOps Foundation)

If you don’t have a copy of the whitepaper yet, you can get a free, open copy here (direct PDF link). We’re working on a digital version where members can access the content and make contributions via GitHub. More updates on that soon.

If you want to join the discussion on FinOps for Kubernetes, join us in our member Slack in the #finops-for-kubernetes channel. We look forward to seeing you there and continuing the dialog.

Why become a member of the FinOps Foundation?

We’re proud to announce that we’re up to about 2700 practitioners, which is an increase from about 1400 just during summer of 2020. This has much to do with joining the Linux Foundation and continuing to appeal to many companies, big and small, who want to make better sense of their cloud cost and usage data.

Becoming a member is free and lets you meet other FinOps practitioners from around the world, and from many different businesses, from startups to the Fortune 100.

Until next time, keep breaking down those silos.