Internal developer platform (IDP)
An internal developer platform is the self-service layer a platform team builds so developers can create, ship and run services without filing tickets or building their own tooling.
An internal developer platform (IDP) is the set of tools, templates and workflows a platform team offers so that developers can create, deploy and operate services themselves. It sits between developers and the underlying infrastructure, and it is built and run as a product with its own users, not as a pile of scripts.
The problem it solves
Picture fifty developers in a dozen teams. Each team needs a cluster, a pipeline, monitoring and secrets, so each team builds its own, a little differently. New joiners learn a different setup per team, fixes don't carry over, and every security or upgrade change has to be made twelve times.
An IDP does that work once and offers it to everyone.
What is usually inside
- A way to start: templates that create a repository with the standard layout.
- Delivery: a pipeline and a deployment mechanism, often GitOps.
- Somewhere to run: a cluster or other runtime, with environments created on request.
- Infrastructure on request: databases, queues and buckets created from a request, not a ticket.
- Observability and security by default: dashboards, alerts, secrets handling and access control on every service.
- A front door: a portal or CLI where developers find all of the above and see what they own.
These are capabilities, not products. Common tools are Backstage for the portal, Kubernetes to run workloads, Terraform or Crossplane for infrastructure, and Argo CD for deployment, but a platform can be built without any of them.
A product, not a project
The CNCF Platforms white paper lists attributes of platforms and of platform teams, and ways to measure success such as user satisfaction and productivity. In practice that means:
- Developers are your customers. Ask what slows them down.
- Start with the smallest thing that removes a real ticket, then grow.
- Measure adoption. If developers route around the platform, the platform is the problem.
How to try it
List every step a new service needs today, from an empty repository to production, and count the tickets and hand-offs. The steps that repeat for every team are your first candidates for the platform. The usual first result is a golden path for the most common kind of service.