Each month, a select group of FinOps Foundation members get together for an Ideas Exchange, an open forum to discuss deep cloud financial management topics. Here are highlights from July’s meeting about building and socializing cloud cost management policies.
If you’re wondering what an Idea Exchange is, here’s a quick reminder.
What is an idea exchange?
It’s an open forum to talk about FinOps challenges, stories, wins, losses, and learnings. Maybe it’s a winning strategy that led to some great outcome. Maybe it was a struggle with a lesson learned. The group comes up with a topic and dives in monthly.
We try to maintain a very “unconference” feel– no planned, dry talks, with a focus on discussion and collaboration. We also get to know one another as there are always new members that join.
After the June 2020 meeting, one of our FinOps Foundation members uploaded a YAML (data file) containing a template for waste sensors on AWS. Check it out, contribute, or comment on the work at the FinOps Foundation GitHub repo.
Creating a usable cloud cost management policy starts on paper
The group warmed up talking about how to create a viable cloud cost management policy. There was definite agreement that you can’t just throw the policy and tooling at teammates. Instead, get the agreement and stakeholdership from users, sponsorship and buy in from leaders, and work with internal communications teams at your business to help spread the word.
Document your policy so teammates have a reference
As FinOps practices grow, so do its lists of definitions, concepts, examples, and more. Each of these looks different depending on the type and size of business you are a part of. The discussion group found that it’s helpful to build a shared doc, microsite, intranet, wiki, and more. The bottom line: create a usable place for people to understand cloud cost management requirements.
These docs can lead to actual practices and tactics, and can give a lot of color and knowledge for cloud teams can self-serve and self-educate. Create small, usable exercises that any team member can start using (e.g., EBS snapshot auditor, scripts that shutdown idle dev EC2 instances in low-risk environments, nights and weekend cloud cleanup policies, etc.). NOTE: Be sure to include ways for teams to explore policy and tooling, but not interrupt mission critical work!
Acclimating teammates to cloud cost governance
Socializing policy utilization and feedback is important as it puts everyone on the same page. The group shared different stories about getting teams to buy into cloud cost management and FinOps policies with everyone agreeing that it starts in dev environments. The main message here is that teams need time to acclimate to new rules and policies around the cloud resources they spin up and use every day.
Ideally, FinOps teams need to give a few notices that policies exist and will be enforced. They also should find clear and open ways to communicate and create exceptions. These exceptions also help make cloud cost management policies a bit more human– knowing that there’s someone around to help exclude a project from shutdown if need be.
Sometimes exceptions are important due to global reasons. Shutting down services overnight makes sense for one timezone, but be mindful if your teams are global. Someone’s downtime might not be everyone’s, this is where detailed exceptions to cloud cost management policies can make the difference between saving money and creating accidental downtime.
In test and production environments (production environments especially), you need to use lessons from dev and increase communication with teammates to ensure you don’t interrupt mission-critical services. Read-only cost policies and settings are important here– the cost of taking down services is high here, and perhaps higher than the amount of cloud cost saved. Take care!
Build a sensible, meaningful account structure
A strong cloud cost management policy is nothing without a thoroughly tagged account structure. Check for tagging, tagging everything missing a tag, and enforcing these policies can improve cost allocation drastically.
Any cloud cost incurred should be billable to a specific team, department, project or workload. And guess what? Even at scale enterprises have trouble here.
FinOps folks can know where to bill costs via this structure rather than forcing teammates to hunt for the right team. Structuring and tagging like this can also help kick the tires of your cloud cost management policy, bring in more teammates to understand the policy, and advocate for policy and tool usage to help keep cloud costs in check.
It’s not always about shutting things off– downsize or autoscale
Shutting off idle resources is one way to save some money. Creating sophisticated policies around autoscaling to still have resources up and running is another way. You’ll want tools to help watch cloud utilization data to ensure that the policy decisions that are made don’t cause bottlenecks or downtime to your business.
If a lower amount of usage tends to trend, this might be an opportunity to downsize the services that your teams might be using.
Team up with security teams
This was an interesting tip from some of the more experienced group members. Lowering cloud costs can sometimes align with hardening security. Removing wasted resources means removing potential attack vectors. Maybe it’s time your FinOps teams hung out with SecOps to see how much more secure and cost efficient yall can make the infrastructure!
Join us next time
After a great discussion, the group voted on August’s topic: Cloud cost budgeting and forecasting.
Idea Exchanges are open to all FinOps Foundation members. If you’d like to be a part of the next one, sign up here. If you’re reading this and aren’t a member yet– what are you waiting for? Sign up to join the FinOps Foundation.
Take care out there and we’ll see you at the next event!