Running `db.collection.getShardDistribution()` when collection has no documents shows `NaN` in several aggregations:
mongos> db.labels_ng_1_1.getShardDistribution() Shard Prod-Shard_2 at Prod-Shard_2/shard:27018 data : 0B docs : 0 chunks : 1 estimated data per chunk : 0B estimated docs per chunk : 0Totals data : 0B docs : 0 chunks : 1 Shard Prod-Shard_2 contains NaN% data, NaN% docs in cluster, avg obj size on shard : NaNGiB