// the find
fluxcd/helm-operator
Successor: https://github.com/fluxcd/helm-controller — The Flux Helm Operator, once upon a time a solution for declarative Helming.
The Flux v1 Helm Operator — a Kubernetes controller that watches HelmRelease CRDs and drives Helm installs/upgrades declaratively. It is explicitly end-of-life and archived; the successor is helm-controller in the flux2 ecosystem. Nobody should be starting a new project on this.
Supported both Helm 2 and Helm 3 simultaneously, with a clean abstraction layer (pkg/helm/v2 and pkg/helm/v3) so the operator could drive either without the user caring. The HelmRelease CRD design — valuesFrom referencing ConfigMaps and Secrets, rollback policy, test hooks — was genuinely ahead of what raw Helm tooling gave you at the time. E2E test suite using bats is thorough and covers rollbacks, CRD handling, and the Helm 2→3 migration path. Prometheus metrics and a Grafana dashboard were included out of the box, not bolted on later.
Dead. Last commit November 2022, officially EOL, and the README tells you to leave. If you are running this in production you have an unpatched Go binary watching your cluster with no upstream fixes coming. The helm-controller replacement has a fundamentally different CRD shape, so migration is a real project, not a flag flip — the valuesFrom API changed, the source separation (HelmRepository/GitRepository as separate objects) is new, and the reconciliation semantics differ. The original design also embedded Helm as a library rather than shelling out, which made Go dependency management painful and limited which Helm version you could use without rebuilding the operator image.