Platform Engineering
Why DevOps Became a Nightmare for Many Developers
Copilot said: The promise of DevOps was compelling: “You build it, you run it.” Developers were given end-to-end responsibility for their code, from writing and deploying it to managing it in production. In theory, this would break down silos and accelerate software delivery. In practice, however, modern cloud-native architectures have created an unsustainable level of cognitive load for software engineers. Today, developers are expected to do far more than write Node.js or Java code. They are also expected to be proficient in Kubernetes, Terraform, CI/CD pipelines, Helm charts, IAM roles, and network security. As a result, the actual process of building product features has slowed considerably. The industry’s response to this challenge is a new discipline: Platform Engineering.
Platform Engineering is not the end of DevOps, but its scalable maturation. It is about building an internal product — the Internal Developer Platform (IDP) — that abstracts away infrastructure complexity, allowing developers to focus on writing business logic
The Internal Developer Platform (IDP) and Backstage
An Internal Developer Platform serves as a bridge between the underlying infrastructure (cloud, Kubernetes, databases) and the developer. It provides self-service capabilities. Instead of having to submit a ticket to the operations team for a new database, or write 500 lines of Terraform code themselves, developers simply go to the internal portal.
One of the most popular open-source frameworks for building an Internal Developer Platform (IDP) is Backstage, originally developed by Spotify and now part of the Cloud Native Computing Foundation (CNCF). Within Backstage, a developer can provision a new microservice with a single click. The platform then automatically generates the Git repository with the appropriate boilerplate code, sets up the CI/CD pipeline in GitHub Actions, configures SonarQube quality checks, and provisions a namespace in the Kubernetes cluster. All of this is done in accordance with corporate standards and security requirements, without requiring the developer to understand the underlying complexity.
Golden Paths and Paved Roads
Within Platform Engineering, the terms “Golden Paths” and “Paved Roads” are frequently used. These refer to predefined, highly recommended technology stacks and workflows established by the platform team that have been proven to be secure, scalable, and fully supported. Developers are free to deviate from these standards—for example, by choosing an unconventional database technology—but doing so means they forfeit the platform’s built-in support, automation, and operational guarantees. If they choose the Golden Path (for example, a stack based on PostgreSQL, React, and Go), deployment, monitoring, security controls, and operational management can be handled almost entirely automatically. This enables development teams to move faster while remaining aligned with organizational standards, best practices, and compliance requirements.
This concept restores balance within the IT organization: the platform team builds the highway (the infrastructure), installs the guardrails (security policies), and the product teams decide what cargo travels along that highway (business features). This approach significantly reduces the onboarding time for new developers and increases standardization across enterprise environments. Developers can become productive faster because they no longer need to learn every infrastructure detail, while organizations benefit from greater consistency, governance, and operational efficiency. Discover the further evolution of software development practices at… AG Connect.
Next: Zero Trust Network Access (ZTNA): Why the Traditional VPN Is Dead
