GH actions gives much more resources and a larger job timeout window. In this way I believe it could fit probably also a KubeCF deployment