Recently, we joined Amazon Web Services (AWS) to talk about one of the fastest-moving transformations in the cloud industry: how Internal Developer Platforms (IDPs) are streamlining modern software delivery. Andy Suderman joined me and Eli Peveto, AWS Partner Solutions Architect to walk through a strategic framework for building an IDP on Amazon Elastic Kubernetes Service (EKS).
During the webinar, we talked through the technical and cultural significance of the discussion, the lessons learned from real-world organizations, and how the combination of open source, automation, and collaboration is enabling organizations to deliver secure, scalable, and developer-friendly infrastructure faster than ever before.
Platform engineering is now essential for organizations aiming to deliver reliable and efficient infrastructure, and IDPs streamline this process as the centerpiece of modern platform teams.
At its core, an internal developer platform is a vehicle for workloads: a structured, efficient, and secure way to deliver code to production. Standardizing deployment workflows creates a “golden path,” removing the friction and risks of manual setup for teams.
By 2026, 80% of large software engineering organizations will establish platform engineering teams as internal providers of reusable services, components and tools for application delivery — up from 45% in 2022.
- Gartner: Unlock Infrastructure Efficiency with Platform Engineering
Some of the major questions today’s enterprise DevOps ecosystems face include:
All of these questions (and more) fit into four key challenge areas we're facing in platform engineering today.
In order to drive true platform adoption and impact, teams must bridge gaps between development, operations, and security. Cross-functional collaboration is essential for establishing clear ownership and coordinating platform initiatives, but legacy silos often slow progress and create resistance to change.
Adoption depends on ease of use. Intuitive tools and robust documentation enable engineers to be productive quickly and to avoid reverting to manual processes.
Platform investments must go beyond technical improvements; they need to visibly advance business goals. This means tying platform metrics directly to organizational outcomes, such as uptime, speed of delivery, or efficiency gains. Clear communication of ROI and continuous measurement are key to sustaining leadership support and ongoing investment.
Today’s platforms operate in increasingly complex, cloud-native ecosystems filled with a multitude of tools, integrations, and scaling needs. Maintaining operational excellence requires careful planning to prevent tool sprawl, controlling technical debt, and streamlining workflows across environments as teams and workloads continue to grow.
When teams maintain separate infrastructure scripts, duplication and gaps in observability are common. AWS’s Platform Strategy sample repo demonstrates how unified templates and workflows solve these issues. To address these interconnected challenges, organizations are turning to modular, open approaches that unify the technical and cultural aspects of platform adoption.
Amazon has collaborated with Adobe, Autodesk, Twilio, and Pfizer, to seek a secure way to manage multi-cluster Kubernetes environments without losing agility.
AWS invested heavily in creating open-source foundations, following the success of CDK EKS Blueprints, which has grown to tens of thousands of weekly downloads. With EKS as the core orchestrator, the platform initiative aims to provide reusable, composable building blocks for organizations adopting platform engineering. This approach allows rapid setup, empowers engineers to self-service infrastructure, and adapts as organizations grow.
With these foundations in place, let’s walk through how a practical internal developer platform can be built and put into action.
At Fairwinds, we’ve specialized in Kubernetes operations for nearly a decade. To address the current need in the market, we created an IDP Quickstart that serves as a simple way organizations can build an IDP using widely supported open-source tooling.
The key components of this IDP include:
Together, these tools create a seamless “golden path” from idea to production. Here’s how it comes together in practice.
Using our Quickstart, you can deploy a Rust backend application from creation to production in minutes. Backstage requests automate infrastructure via IaC, with Argo CD ensuring changes flow straight to workloads. Multi-account or multi-cluster setups integrate with AWS Organizations and Landing Zone, allowing network isolation and resource governance across business units. Policy-as-code with Open Policy Agent (OPA), Polaris, or Kyverno enforces compliance at every step. While launching a developer platform is the first step, long-term value depends on reliable day-2 operations and constant refinement.
Building an IDP on EKS is only the beginning; operational success is defined by how well you handle upgrades, security, tenancy, and disaster recovery. Teams must plan for day-2 operations from the start to avoid downtime and security gaps.
Regular cluster and workload upgrades require careful coordination: upgrade dev and staging clusters first to allow for soak time before pushing to production, leverage blue-green deployments for major version bumps, and always validate compatibility for add-ons before controller upgrades.
Namespace-level tenancy via Kubernetes Role-Based Access Controls (RBAC) and network policies, routine secrets rotation, and pod disruption budgets all enforce secure boundaries. For stateful apps, regular Velero and RDS backups are key aspects of disaster recovery.
Once the foundations of operational excellence are in place, the focus shifts to monitoring, alerting, and measuring success at scale.
Scalable monitoring and alerting are crucial for recognizing issues early and maintaining Service Level Objectives (SLOs). Integrating OpenTelemetry Software Development Kits (SDKs) into workloads enables standardized traces and metrics, which can flow to AWS CloudWatch, Prometheus, or any number of backends. Automated, reasonable alert defaults can be deployed for pod, node, ingress, and workload health. DORA metrics (deployment frequency, lead time, recovery time, change failure rate) can be surfaced through custom pipelines so teams can quantify delivery performance over time. Use community operators and managed add-ons for rapid deployment of monitoring infrastructure.
Observability and alerting helps teams monitor not only Kubernetes-native but also hybrid workloads, therefore the platform must allow flexibility for hybrid and legacy systems as well.
Not every application fits the Kubernetes model. EKS-based IDPs should provide escape hatches for integrating workloads running on Amazon Elastic Container Service (ECS) or Lambda using tools like DreamFactory alongside microservices. Documented patterns for securely connecting non-K8s services, managing hybrid network policies, and synchronized monitoring give platform teams the flexibility they need to support business needs without forcing a “one size fits all” approach.
By focusing on both technical workflows and real developer needs, teams can realize measurable productivity gains and smoother onboarding.
The design principles in IDP Quickstart for EKS simplify developer onboarding, so new engineers can deploy production‑ready apps after just three hours, compared to the 16 hours of environment setup training required previously. Standardized practices prevent recurring CI/CD errors, so new developers can hit the ground running. This not only enables “shift-left” adoption but also prevents repetitive CI/CD errors that previously slowed teams down.
If fragmented pipelines and slow onboarding have hindered your team, this webinar makes one thing clear: fast, scalable, and adaptable IDP adoption is achievable now. The tools and best practices empower organizations to accelerate delivery, improve developer satisfaction, and maintain robust governance. Your journey to platform engineering success can start today.
Watch the full webinar on demand: Platform Adoption in Months Instead of Years.
Photo by taopaodao on Unsplash.