This might require a change on kubedog, or even potentially some way to have a background job wait and terminate the kubedog tail after N seconds...