Platform Engineering: The Foundation for shifting FinOps left

Platform Engineering: The Foundation for shifting FinOps left

FinOps is currently practiced in most organizations as a downstream control mechanism to manage capacity reservations or respond to cost anomalies. Proactive measures typically intervene at the software architecture level. Platform Engineering creates an ideal starting point for integrating FinOps instrumentation and governance tools across the entire software landscape through the centralization and standardization of workflows.

What’s Behind Platform Engineering

Platform Engineering is a discipline that centralizes existing knowledge and standards around software development processes within organizations and makes them visible and accessible to all teams via self-service. Goals include reducing cognitive load for development teams so they can focus on business logic, as well as accelerating the development of new features by minimizing undifferentiated heavy lifting within the organization.

Platform Engineering explicitly does not aim to impose a one-size-fits-all silver bullet standard solution on all teams, but rather to consolidate best practices into so-called Golden Paths, that allow development teams to deviate if the cost making part of the platform their responsibility provides a tangible benefit. These Golden Paths are compiled in a Service Catalog and made visible and accessible to development teams in the form of an Internal Developer Platform managed by the platform team.

How FinOps and Platform Engineering Benefit Each Other

Platform Engineering creates a technical foundation for anchoring FinOps practices early in the lifecycle of software products, thereby providing teams with tools for autonomous cost control from the start.

Platform Standards for Better Visibility

A well-designed platform unifies the paths to target environments - whether hyperscalers, Kubernetes, or simply your own data center. This enables a consistent and efficient tagging and metadata strategy for every provisioned resource. The team selects the Golden Path and cost centers, responsibilities, environments, and other important information are attached to all resources by the platform - even across cloud providers.

This enables the platform team to provide development teams with an always up-to-date dashboard containing itemized billing data for their used services, across multiple clouds, SaaS providers, or internal data centers.

Automated Governance over Manual Approval Processes

Software development requires speed. This doesn’t align with lengthy approval processes by traditional accounting departments and committees. In a central platform, projects can be automatically approved based on known budget limits, only overruns need to be approved by the necessary stakeholders - and the platform can automatically trigger the approval requests.

Small Optimization, Big Leverage

Through a central platform, even tiny changes can have a large impact simply by being rolled out across many products in the corporate landscape at once: lifecycle policies for storage, automatic shutdown of test environments outside business hours, or new, cheaper instance types for VMs.

Clean Observability Everywhere Also Helps Cost Transparency

With observability solutions rolled out uniformly across services, FinOps-relevant signals can also be deployed, improved, and evaluated comprehensively.

• Unit Economics: Through the platform, systems for measuring cost-per-request or cost-per-transaction can be rolled out comprehensively, giving the FinOps team direct insight into resource efficiency.
• Correlation of Performance and Resource Efficiency: An undersized resource at one point in the system can form a bottleneck that slows down the entire system to the point where overall costs are driven up by longer idle downstream resources. Consistently rolled-out monitoring can correlate these bottlenecks with the total costs of the system, providing a better overall picture of the software landscape.
• Identify Consistently Unused Capacity: With good observability, the utilization of ordered resources can be measured in detail and adjusted by the development team itself. Low utilization of resources provided by the platform can be converted by the team into adjusted sizing and thus realized savings at the organizational level.

The Standardized CI/CD Pipeline as Platform Orchestrator

Most of these functions can already be implemented today with the CI/CD solution found in most companies, such as GitLab CI, GitHub Actions, or Azure DevOps. Maintained by the platform team, a CI/CD pipeline can handle numerous standard compliance and governance tasks for all teams.

Automated Policy Gates

The CI pipeline can act as a gatekeeper while also keeping FinOps goals in mind.

• Budget Checks: Tools like Infracost are able to estimate cloud costs before any rollout occurs. This allows the organization to gate deployments that exceed team budgets to gated behind stakeholder approval.
• Rightsizing Checks: Deployments that request oversized resources, or generate suspicious cost differences can be held back for additional approval.
• Access Rights: Automatically ensure before deployment that IAM policies are set as precisely as possible and prevent granting unnecessarily broad access rights.

Environment-Specific Optimizations

With standardized rollout processes for development environments, sizing for development environments can be performed in the pipeline.

• Reduced Instance Types: In development environments, instance types for various resources can be automatically be downsized using predeterined strategies. Depending on the situation, the platform can even decide to share database instances between multiple deployments, deploy databases inside containers instead of full managed instances.
• Disabling Redundancies: High-availability measures such as redundant deployments or multi-AZ failover database setups are rarely necessary for development environments. The pipeline can ensure that deployments are made without redundancy there.
• Time-Controlled Shutdowns: Automated jobs to shut down development environments at the end of the business day and spin them back up in the morning before work begins can be deployed directly via the pipeline.

Optimization of the Pipeline Itself

By instrumenting the CI pipeline itself, the platform team can measure its resource usage and pinpoint inefficient processes, such as long-waiting instances, for optimization. These improvements commonly also help shorten pipeline runtime, thereby reducing cycle times for development teams.

Conclusion: The Organization Benefits When FinOps Is Considered in Platform Engineering

Proactive integration of FinOps thinking and processes at the platform level benefits the organization on multiple levels.

• FinOps teams gain comprehensive insight into the organization’s resource efficiency and can automate FinOps governance processes within the platform. This enables FinOps teams to act proactively rather than just reacting to existing issues.
• Platform teams receive data prepared by FinOps teams and can thus optimize resource efficiency on a large scale at the platform level.
• Development teams receive shortcuts to lengthy processes via Golden Paths augmented with FinOps solutions, as well as overview and necessary information about the costs and efficiency of the systems under their management.

Therefore, it’s worthwhile on multiple levels to consider FinOps in the company’s platform strategy.

You want to anchor FinOps principles directly into your platform strategy? We are ready to support you in making cost efficiency an essential part of your Internal Developer Platform - from tagging strategy, to automated budget governance. Let us find out which quick wins you can realize in your organization!