Trends in Cloud FinOps implementation
By Magesh Kasthuri
IT spend optimization is one of the primary goals for any organization. Cloud service usage for computing, storage, and the network becomes uncontrollable as its cloud footprint grows larger. At this stage, controlling and managing the cost of cloud usage becomes essential. In cloud terminology, this is called FinOps. FinOps provides a view between price (incurred usage) and value (ideal use). When it comes to cloud adoption, Return on Investment (ROI) in cloud transformation and Total Cost to Organization (TCO) in cloud cost spent is key metrics being looked at by business stakeholders and CxOs.
FinOps includes various disciplines, such as business analysis, cloud engineering, cloud automation, financial operations, data analytics, and operational re-engineering. Together, these business processes work in a single interoperative flow.
The FinOps lifecycle starts with a design to define cost metrics, which are mapped to business metrics, and technical or usage metrics. Business metrics are a set of parameters that help the stakeholders get a high-level view of cloud cost expense, including CAPEX and OPEX in which CAPEX is a combination of one-time sunk cost, integration cost, and locked-in cost. CAPEX is used to calculate the ROI in cloud investment. Both CAPEX and OPEX are crucial to understanding the Total Cost to Organization (TCO) which involves all direct and indirect costs.
The two stages of FinOps operations are:
Stage 1
Resource tagging to bring a uniform pattern of resource identification across the cloud organization and
Stage 2
Metric evaluations can pull the usage metrics and prepare usage reports for application owners and projection calculations for business owners.
We need to consider the upfront cost (invested in a public and private cloud).
The State of FinOps 2021 report states that only 15% of organizations have mature cloud cost management. In 2022, many enterprises aim to implement FinOps and cloud cost management for multi-cloud platforms to have cloud-agnostic FinOps implementation. An example of common patterns for FinOps implementation is an auto-scaling Kubernetes instance.
The goal here is to collect metrics of usage patterns from Kubernetes pod clusters. Since there are native implementations of Kubernetes in Azure (AKS), Amazon (EKS/Fargate), and Google (GKE), implementing a cloud-agnostic horizontal pod auto-scaler can be cost-effective during fan-in and fan-out of instances or nodes.
Here are two key steps to implementing FinOps for auto-scaling Kubernetes instances to enable efficient auto-scaling and cost-optimized Kubernetes clusters.
Step 1: Define threshold limits for metrics like CPU > 80%, and, based on those limits, define rules for Pod auto-scaler.
Step 2: Based on threshold rules developed in Step 1, configure horizontal auto-scaler to enable the creation of new pods to Kubernetes clusters automatically.
FinOps examples
In a cloud setup for a banking customer, we developed a FinOps framework that includes
- FinOps governance.
- Workflow-based infra-provisioning in a multi-cloud environment integrated with ITSM ticketing.
- A maturity model for FinOps delivery.
- Federated cloud cost management for different lines of business,
- Resource tagging strategy and tagging policy guidelines,
- Technical and business metrics for cloud cost management dashboard and reporting,
- Utilization reports are filtered by resource tags like a cost center and environment or stage.
After the initial setup, we enhanced the implementation to include data analytics from various feeds for resource utilization, usage pattern analysis, automated resource management for optimized resource usage, integration for cloud cost management, and a target operating model for showback or chargeback.
In a second example, we set up a cloud business office focusing on cloud governance, cloud migration factory setup, and cloud cost management framework implementation. The focus of the initial phase of setting up the cloud business office was to implement cost optimization for EKS container instances in the AWS platform, including cost transparency services like the Wall of Optimum utility and FinOps Dashboard, cost management services like budget actions, alert notifications, tagging hygiene, savings plan, right-sizing VM instances, migrating to Spot/Reserved instances and FinOps reports.
In both examples, client organizations realized immediate benefits from the visibility they gained into the technical and business metrics on cloud usage provided by our FinOps implementation.
