a customer is seeing 5xx errors on the /@{user}/{workspace_and_agent}/apps/{workspaceapp}/* route after the workspace is turned off. what is the cause of this, is the coder_app healthcheck continuing to be hit after the workspace is stopped?