You’ve got Drone pipelines running like clockwork, then someone whispers “Kustomize” and suddenly you’re questioning everything. Should you bake your Kubernetes overlays into CI? Should each environment have its own manifest tweaks? Drone Kustomize exists for exactly that reason, and it can save you from an endless patch spaghetti.
Drone handles build, test, and deploy automation. Kustomize manages modular Kubernetes manifests without copying YAML fifty times. Alone, they’re useful. Together, they turn deploy pipelines into version-controlled configuration systems that don’t crumble the moment staging needs a different image tag.
When you run Drone Kustomize, your pipeline can dynamically overlay Kubernetes manifests at build time. Instead of managing separate manifests per cluster or team, Kustomize patches a base set with the right settings for that environment. Drone picks up those versions, applies credentials through Kubernetes service accounts or OIDC bindings, and rolls out updates safely. Think of it as declarative automation meeting declarative infrastructure.
The workflow looks like this: Drone executes a pipeline step, pulls the repo containing your manifests, runs kustomize build on the intended overlay, then applies it through your preferred context. Identity can come from AWS IAM roles, GCP Workload Identity, or even your internal OIDC provider. Permissions stay least-privileged. Each change is logged in Drone, and every cluster configuration remains consistent.
When troubleshooting, start with the basics: confirm your path mappings and base directories in the Drone step. Most failed builds trace back to missing kustomization.yaml references or absent patches. Treat overlays as immutable — never tinker mid-pipeline. If your team uses RBAC, map service accounts properly between namespaces so Kustomize-generated manifests don’t fail deployment post-render.