UserWarning for specified Dask chunks when reading Zarr from S3 bucket

[mliukis]

When reading Zarr store from S3 bucket:

import s3fs
import xarray as xr

s3 = s3fs.S3FileSystem(anon=True)
zarr_store = s3fs.S3Map(root=cube_url, s3=s3_in, check=False)
with xr.open_dataset(zarr_store, decode_timedelta=False, engine='zarr', consolidated=True, chunks={'mid_date': 250}) as ds:
   ...

I am getting the following user warning:

..../site-packages/xarray/core/dataset.py:410: UserWarning: Specified Dask chunks (250, 250, 250, 250, 250, 194) would separate on disks chunk shape 361 for dimension mid_date. This could degrade performance. Consider rechunking after loading instead.

I don't get this warning when I open the same Zarr store from the local file system. Does anybody know why I am getting this warning when Zarr is stored in S3 bucket?

I need to make sure that I open existing Zarr with Dask as I need to stream write it to the NetCDF file later on in the processing.