The Temporal Worker Controller now supports autoscaling for versioned Worker Deployments on Kubernetes via a new `WorkerResourceTemplate` custom resource. Because Temporal's durable execution model requires running multiple Worker versions simultaneously (rainbow deployments), each version needs independent scaling. WorkerResourceTemplate lets you define an HPA or PodDisruptionBudget once, and the controller automatically creates a correctly targeted, labeled instance per Worker version, injects the right scaleTargetRef, and cleans up when a version is retired. Scaling can be driven by CPU, memory, or external metrics like Temporal's `approximate_backlog_count` labeled by Worker version. Security guardrails include a validating webhook with RBAC SubjectAccessReviews and blocked workload resource kinds. Available in Helm chart v0.23.0 / appVersion v1.5.0.
