Skip to content
Back to glossary
DevOps Methodology

What is Platform Engineering?

Platform Engineering builds Internal Developer Platforms (IDPs) that industrialise the workflows of development teams: provisioning, deployment, observability, security.

What Platform Engineering does

Platform Engineering is an approach where internal infrastructure is treated as a product targeting the company's own development teams. The core idea: instead of asking every developer to understand Kubernetes, Terraform, ArgoCD, monitoring, TLS certificates, secret management and network policies, a "platform" team builds an Internal Developer Platform (IDP) that abstracts those complexities behind simple interfaces.

Concretely, a developer no longer deploys an application by writing Kubernetes manifests: they push code, and a standardised workflow builds the image, deploys it, provisions certificates, configures monitoring and exposes the service. The platform is documented, versioned, and treated with the same rigour as an external product.

Why adopt Platform Engineering

Reduce developer cognitive load. According to the Puppet State of DevOps 2024 report, engineers spend on average 20 to 30% of their time on infrastructure tasks they do not fully master. A good IDP cuts this friction and redirects that energy to the product.

Standardise best practices. Security, compliance (nFADP, GDPR), observability and cost-management policies become built-in to the platform. No need to verify that every team applies the right scanning, quotas, networks: the platform enforces them by construction.

Faster onboarding. A new engineer ships their first production change in days, versus weeks in an environment where they must master the entire DevOps stack.

Offload the DevOps team. Without a platform, the DevOps team becomes a bottleneck: every deployment, every new service, every audit goes through them. An IDP turns the DevOps team into platform builders rather than ticket executors.

Typical components of an IDP

On Hidora engagements, a mature internal platform typically combines:

  • Self-service provisioning: Backstage, Port or a custom interface to create a new service with its dependencies (database, cache, queue).
  • Standard templates: application skeletons with CI/CD, observability, monitoring and secrets pre-wired.
  • GitOps workflow: automated deployment via ArgoCD or Flux, no manual intervention.
  • Centralised observability: Prometheus, Grafana, Loki accessible to everyone, with per-service dashboards generated automatically.
  • Central secret management: Vault, External Secrets Operator so no secret ever lives in Git.
  • Automatic security policies: OPA, Kyverno to validate deployments against organisational standards.

When to invest in a platform

Platform Engineering becomes worthwhile when you have at least 4-5 product teams as consumers, or around 30 engineers. Below that, the build effort exceeds the gains. For a Swiss SME of 50 people with a single product team, a documented standard DevOps stack is enough.

When Platform Engineering fails

The most frequent trap: building a platform without consulting product teams. A top-down IDP, without co-design, ends up ignored. The right reflex is to treat developers as customers: interviews, adoption metrics, continuous iteration, internal NPS.

Related Hidora services