To get the most from cloud investments as your business grows, FinOps has to grow with you. That means weaving automation into the fabric of core FinOps practices: rate optimization and workload optimization.
This blog explores the FinOps Automation Framework—a three-phase journey that turns manual cost management into a streamlined, automated process driven by workloads and business value.
Phase 1: Automating commitment management
Traditionally, Reserved Instances (RIs) and Savings Plans (SPs) are seen as mechanisms to reduce costs for predictable workloads. However, correct automation also can make them fit for variable workloads, with two key strategies:
- Micro-purchasing of rigid SPs or RIs (a.k.a. layering or stacking)
- Introducing convertible RIs and actively using them to cover usage spikes (a.k.a. seeding and expansion)
Automating the lifecycle of convertible RIs makes commitments relevant in dynamic environments such as K8s and spikey workloads.
Automating the commitment lifecycle involves:
- Continuously analyzing usage patterns to forecast demand
- Executing RI/SP/CUD purchases and renewals (or deciding against renewals)
- Modifying convertible RI volume and terms to eliminate over-commitment
The resulting commitment blend will balance fully utilized long-term, rigid commitments (highest discounts) with burstable, alterable commitments (slightly lower discounts). This means two distinct benefits for cloud cost reduction:
- Increased coverage of your cloud workloads by discounted resources, meaning pricey on-demand consumption is reduced
- Full utilization of purchased commitments, to ensure effective savings rate (ESR) is almost equal to the commitments’ average discount rate
Interested in more extensive optimization? Explore more more strategies for maximizing cloud commitments in this Flexera blog post.
Phase 2: Automating spot instances
In the second phase, we expand automation to include the purchasing and graceful replacement of spot instances. Most organizations experiment with these deeply discounted compute resources on stateless workloads in their Dev and QA environments. However, with the right tooling, spot instances can be used for stateful applications and in production, too.
With both inventories (commitments and spot instances) auto-managed, we can now configure “Dual Autoscaling.” This means that, for each autoscaling event, automation intelligently selects between existing commitments and spot instances based on workload characteristics, availability and pricing.
This phase marks the completion of rate optimization automation. You can evaluate its financial benefit by a “rate optimization ESR”—a modification of the original ESR that divides commitment and spot instance spend by the alternative cost of the resources consumed in fact.
Rate optimization ESR = Commitment spend + spot instances spend━━━━━━━━━━━━━━━━━On-demand equivalent of actual usage
You can further explore how automated spot instance management and dual autoscaling drive smarter cloud optimization in this Flexera blog entry on new auto-upgrades and commitment autoscaling releases for our Ocean customers that are Azure users.
Phase 3: Automating workload optimization
With rate optimization fully automated, the final phase targets workload optimization. This involves eliminating idle capacity and improving resource utilization through techniques such as:
- Rightsizing—automatically adjusting resource sizes to match actual usage, so the same job can be done with smaller machines
- Bin packing—consolidating workloads to reduce fragmentation and improve density, so the same jobs can be done with fewer machines
- Configuring dynamic storage volumes—scaling storage up and down based on real-time demand
- Shutdown schedules—powering down non-essential resources during off-hours
Automating these practices ensures that cloud environments remain lean and efficient, continuously aligning infrastructure with business needs.
You can dive deeper into workload optimization in these posts:
- FinOps as Code: Because you can’t git revert your cloud bill
- The practical FinOps roadmap series: Optimizing cloud usage and rates
Wrapping up: How can all this work together?
Imagine a cloud environment where rate optimization and workload optimization are seamlessly integrated—no longer siloed, but working in concert to maximize value.
In this ideal state, workload optimization is managed natively through Infrastructure as Code (IaC) tools such as Terraform. IaC management enables dynamic, policy-driven resource adjustments that keep infrastructure (and DevOps tooling stack) lean and efficient.
On the FinOps side, all optimization efforts, from commitments to autoscaling, are continuously measured and visualized in your cloud financial management tool. This can give you real-time insight into savings, efficiency and business impact.
This is the future of FinOps: fully automated, natively integrated, and always aligned with your business goals. Ready to take the next step? Contact us to accelerate your journey.
