A Helm Chart for `workerd` to run it on Kubernetes clusters easily with some documentation, would make large scale deployments far easier to approach.