I originally mirrored the params for the dataset and dataverse searches, but that seemed to throttle the dataverse one to only one (seemingly random) dataverse, and this was fixed by simply removing 'fq': status; unclear why this doesn't work since this seems to work fine in a web call using that filter.