Learn Backstage - Modern Ecosystem & Final Reflection

#Introduction

Episode 21 brought Backstage to production and set up adoption. Episode 22 is the final episode of the Learn Backstage series — and at the same time a window into the future. You'll see Backstage's modern features in 2026, compare them with alternatives in the IDP market, recap the whole journey from episodes 0 to 21, and close with a production-grade checklist plus a look at the future of platform engineering.

#Modern Backstage Features in 2026

#New Frontend System as Default

In episode 17 you met the New Frontend System. Since 2026, this system has become the default in Backstage: frontend plugins are built with createFrontendPlugin, extension points, and dependency injection as the standard patterns. The experience for plugin developers is now uniform and easier to learn.

#React 19 and TypeScript Modernization

The Backstage frontend keeps being modernized. The migration to React 19 brings a more efficient rendering model, while the TypeScript codebase is updated to use the latest language features. For you, this means the plugin code you write today aligns with the direction of future development.

#MCP Integration for AI Agents

One of the most interesting features in 2026 is MCP (Model Context Protocol) integration. Backstage now provides an MCP server that lets AI agents talk to the platform over the Streamable HTTP protocol. Through MCP, an AI agent can:

• Search for entities in the software catalog.
• Run the scaffolder to create a new service from a template.
• Answer questions about service status and ownership.

The MCP server exposes actions registered in the backend, so agents can use the platform's capabilities without opening a manual interface. On the client side, server registration is done through ordinary MCP configuration:

{
  "mcpServers": {
    "backstage-actions": {
      "url": "http://backstage.example.com/api/mcp-actions/v1",
      "headers": {
        "Authorization": "Bearer ${MCP_TOKEN}"
      }
    }
  }
}

@backstage/backend-plugin-api remains the basis for development, and MCP opens the door to new workflows — developers can create services through natural language commands, and agents can navigate the catalog without a manual interface.

#Catalog Model Extensibility

Backstage's catalog model isn't rigid. You can define custom kinds, add processors to ingest new data sources, and build relations between entities that fit your organization's domain. This extensibility keeps the catalog relevant even as the organization grows and changes.

#Four Keys Telemetry

Inspired by DORA metrics, Four Keys measures software delivery performance through four numbers. By combining analytics events (episode 20) with CI/CD data, Backstage helps teams see whether process improvements are actually working.

#Backstage vs the Alternatives

Backstage isn't the only player. Understanding the comparison helps you make the right decision:

#When to Choose Backstage

Backstage wins when an organization wants control and flexibility: a large plugin ecosystem, deep customization possibilities, and no vendor lock-in. It's best suited to organizations that have a platform team capable of maintaining and adapting the platform themselves.

A SaaS IDP becomes more attractive when speed is everything — a small team wants a portal running fast, without the burden of maintaining the catalog and upgrades. This decision isn't about right or wrong, but about team capacity and long-term needs.

#Journey Recap: Episodes 0-21

Let's pause and look back. Across 22 episodes you built a complete understanding of Backstage:

• Episodes 0-3 — the foundation: prerequisites, history, architecture concepts, and initial application bootstrap.
• Episodes 4-5 — the software catalog: basic entities, ingestion, and processing.
• Episodes 6-7 — basic scaffolder and TechDocs.
• Episodes 8-9 — authentication, identity, and advanced configuration.
• Episodes 10-12 — advanced scaffolder, catalog data model, and persistence.
• Episodes 13-16 — RBAC, secrets and external integration, deployment hardening, and search.
• Episodes 17-18 — frontend and backend plugin development.
• Episodes 19-21 — scaling, observability, production deployment, and adoption.

From the first catalog to a platform serving thousands of engineers — all the pieces connect into one system.

#The Production-Grade Backstage Checklist

As a technical close, here's a checklist that brings together the entire series:

• Catalog quality — entities have a clear owner, description, documentation, and technology; no orphan entities.
• Security — authentication active, authorization and RBAC configured, secrets managed securely, dependencies always updated.
• Observability — structured logs, metrics and alerts active, traces available when needed.
• Upgrade hygiene — weekly releases followed, backstage-cli versions:bump run regularly, and staging tests before production.

A platform that meets this checklist is ready to be relied on by the organization.

#The Future of Platform Engineering

Platform engineering is becoming a core practice in many organizations: the platform is no longer a collection of tools, but a product with users and a roadmap. The direction is clear — increasingly automated golden paths, an increasingly living catalog, and the presence of AI agents using the platform through MCP. Backstage, with its open-source ecosystem, sits at the center of this shift.

#Conclusion

Episode 22 closes the Learn Backstage series. You've completed a full journey from episodes 0 to 21: foundations and architecture, the software catalog, scaffolder, TechDocs, authentication and RBAC, secrets, search, frontend and backend plugin development, scaling, observability, production deployment, and adoption strategies. In this final episode, you also got to know Backstage's modern 2026 ecosystem and how to position it among IDP alternatives.

The key takeaways:

• Backstage is a platform, not a finished product — its value grows from the catalog, golden paths, and plugin ecosystem you build.
• Flexibility is both strength and cost — choose Backstage when you have the team capacity to maintain it.
• Operational discipline underpins everything — catalog quality, security, observability, and upgrade hygiene are the four production-grade pillars.
• Platform engineering keeps evolving — golden paths, catalogs, and AI agents will determine how developers work going forward.

Thank you for following the entire series. The skills you've built — understanding, customizing, and operating a developer portal — are a solid foundation for continuing in devops and other platform engineering topics. Your journey in the platform world doesn't stop here; I hope this material serves as a stepping stone to explore the next practices. Happy building your own platform!