Skip to content
EgyKodeAcademy
Advanced40 min

The Future (Platform Engineering)

After this chapter you can

  • Explain the golden path and why platforms beat ticket queues

Introduction to Platform Engineering#

For the last 15 years, the industry buzzword was "DevOps". The idea was that developers and operations should be the same team. A developer writes the Java code, writes the Dockerfile, writes the Terraform, deploys the servers, and wakes up at 2 AM when it crashes.

We learned a hard lesson: Cognitive Overload. You cannot expect a frontend React developer to also be an expert in AWS VPC Routing, Kubernetes eBPF network policies, and HashiCorp Vault token rotation. It is too much to learn.

Platform Engineering is the evolution of DevOps.


Level 1 — Beginner#

What is Platform Engineering?#

Imagine you are a chef in a restaurant (A Developer).

  • The DevOps Way: The chef has to build the stove, plumb the gas lines, install the fire alarms, and then cook the steak. The chef is exhausted and the steak is burnt.
  • The Platform Engineering Way: A dedicated team of engineers builds a pristine, safe, commercial kitchen. They give the chef a single button that says "Turn on Stove." The chef cooks a perfect steak.

Platform Engineers do not build the product (the website). They build the Platform that the developers use to build the product.

The Internal Developer Platform (IDP)#

An IDP is a self-service portal (like an internal Heroku). A developer logs into a website, clicks "Create New Microservice", and types billing-api. In the background, the Platform automatically:

  1. Creates a GitHub repository.
  2. Creates the Jenkins CI pipeline.
  3. Provisions a PostgreSQL database in AWS using Terraform.
  4. Generates the Kubernetes YAML files.
  5. Injects the monitoring dashboards into Grafana.

The developer starts writing Java code on Day 1.


Level 2 — Intermediate#

Treating the Platform as a Product#

In traditional IT, the Operations team acted as a ticketing system. ("Please open a Jira ticket and we will provision your database in 3 weeks.") Platform Engineering treats the Platform as a Product. The developers are the customers. If the Platform is too hard to use, or if the documentation is bad, the developers will hate it. A good Platform Team conducts user interviews, measures adoption rates, and continuously improves the UX (User Experience) of the platform.

Golden Paths (Paved Roads)#

You don't force developers to use the platform. You make the platform so easy to use that they choose it. This is called the Golden Path.

  • The Golden Path: You use the standard Java Spring Boot template. The Platform team handles all CI/CD, security scanning, scaling, and on-call paging for you automatically.
  • Going Off-Road: You decide to write your microservice in Haskell using a custom CI system. That's fine, but you are responsible for the security audits, and you are on-call at 2 AM.

Level 3 — Advanced#

Analyzing the Architecture (Backstage)#

How do you actually build an Internal Developer Platform? You use Backstage (an open-source framework created by Spotify).

Backstage has a feature called Software Templates. When a developer clicks "Create", Backstage executes a declarative YAML file.

yaml
apiVersion: scaffolder.backstage.io/v1beta3
kind: Template
metadata:
  name: spring-boot-microservice
  title: Java Spring Boot API
spec:
  steps:
    - id: fetch-template
      name: Fetch Skeleton Code
      action: fetch:template
      input:
        url: ./skeletons/spring-boot
        
    - id: publish-github
      name: Create GitHub Repository
      action: publish:github
      input:
        repoUrl: github.com?owner=ivolve&repo=${{ parameters.name }}
        
    - id: trigger-argocd
      name: Register with ArgoCD
      action: argocd:create-app

Line-by-Line Breakdown:

  • action: fetch:template: Backstage downloads a pristine, highly-secure, pre-approved Java skeleton codebase.
  • action: publish:github: It uses the GitHub API to dynamically create a brand new repository and pushes the skeleton code to it.
  • action: argocd:create-app: It natively talks to ArgoCD to register the new repository. The CI/CD pipeline is fully wired up in 5 seconds.

Level 4 — Enterprise#

Developer Portals and Software Catalogs#

In an enterprise with 2,000 microservices, if a database goes down, how do you know which team owns the database? You use a Software Catalog (the core feature of Backstage).

Every single microservice in the company contains a catalog-info.yaml file in its Git repository.

yaml
apiVersion: backstage.io/v1alpha1
kind: Component
metadata:
  name: payment-api
  description: Processes credit cards
spec:
  type: service
  owner: team-billing
  lifecycle: production
  dependsOn: ['resource:rds-payment-db']

Backstage ingests these files and builds a visual graph of the entire company. You can click on the payment-api, see that it is owned by team-billing, click a link to view their Grafana dashboard, click a link to view their Jenkins pipeline, and click a link to page their on-call engineer in PagerDuty. It ends the chaos of distributed microservice sprawl.

Cognitive Load and Abstraction#

The goal of Platform Engineering is to find the perfect level of abstraction.

  • Too little abstraction: You force developers to write raw Kubernetes YAML. They make mistakes and introduce security vulnerabilities.
  • Too much abstraction: You hide everything behind a GUI. When the system breaks, the developers have no idea how to troubleshoot it because they don't understand the underlying technology.

The best platforms use Progressive Disclosure. They provide a simple 5-line values.yaml file by default (The Golden Path). But if an advanced developer needs to add a custom Istio VirtualService, the platform allows them to "eject" and write the raw YAML safely.


Interview Questions#

Beginner#

Q: What is the main difference between DevOps and Platform Engineering? A: DevOps is a culture where Dev and Ops work closely together (often combining the roles). Platform Engineering is a specialized discipline where a dedicated team builds a self-service product (the platform) for the developers to consume, reducing the cognitive load on the developers.

Intermediate#

Q: What is an Internal Developer Platform (IDP)? A: An IDP is a self-service portal (often built with tools like Backstage) that standardizes and automates the creation of infrastructure, CI/CD pipelines, and application scaffolding. It allows developers to spin up new, fully-compliant microservices in minutes without needing to understand Terraform or Kubernetes.

Senior#

Q: Explain the concept of the "Golden Path" (or Paved Road) and why it is better than strict IT mandates. A: Strict IT mandates force developers to use specific tools, causing friction and slowing down innovation. A Golden Path is a fully supported, highly automated, secure route provided by the Platform team. Developers are technically allowed to choose different tools (going off-road), but they lose the automation, support, and on-call protection of the Golden Path. Ultimately, 95% of developers will voluntarily choose the Golden Path simply because it is the easiest way to get their job done.

Principal/Architect#

Q: You are rolling out Backstage as an IDP to 500 developers. Currently, developers deploy infrastructure by opening a Jira ticket to the Ops team. How do you architect the transition to ensure the IDP actually gets adopted, rather than becoming another unused corporate tool? A: Adoption requires treating the Platform as a Product.

  1. User Research: Do not build what you think is cool. Interview the developers. Find their biggest pain point (e.g., "It takes 3 weeks to get a database").
  2. Minimum Viable Product (MVP): Build a template in Backstage that provisions a database in 5 minutes via Terraform and Crossplane.
  3. Evangelism: Onboard one highly influential "lighthouse" development team. Let them experience the 5-minute database. They will brag about it to other teams.
  4. Metrics: Measure the reduction in lead time (DORA metrics). Present this data to leadership to secure funding to expand the platform to cover CI/CD and Kubernetes deployments. Contents | The Worst Day (Disaster Recovery) |

Check yourself

4 questions from this chapter. Try answering before you look.

  • What is the main difference between DevOps and Platform Engineering?
  • What is an Internal Developer Platform (IDP)?
  • Explain the concept of the "Golden Path" (or Paved Road) and why it is better than strict IT mandates.
  • You are rolling out Backstage as an IDP to 500 developers. Currently, developers deploy infrastructure by opening a Jira ticket to the Ops team. How do you architect the transition to ensure the IDP actually gets adopted, rather than becoming another unused corporate tool?
Questions from the curriculum →

Recommended free courses

All courses

Another way to learn this — external, free, and not affiliated with EgyKode.