KeyError when selecting "nearest" data with given tolerance

What happened: Selecting data with ds.sel(index=given_index, method="nearest", tolerance=tolerance) only works in case for each given_index exists an index that is within the given tolerance, otherwise a `KeyError: "not all values found in index 'index'" occurs.

What you expected to happen: I would expect, that those indices that are not within the tolerance would simply be dropped.

Minimal Complete Verifiable Example:

import xarray as xr
ds = xr.DataArray([1,2,3,4,5], dims=["lat"], coords={'lat':[10,20,30,50,60]})

# working example, all latitudes are within the tolerance
ds.sel(lat=[5,15,40], method="nearest", tolerance=10)
#<xarray.DataArray (lat: 3)>
#array([1, 2, 4])
#Coordinates:
#  * lat      (lat) int64 10 20 50

# failing example, latitude 40 is not within the tolerance
ds.sel(lat=[5,15,40], method="nearest", tolerance=5)
# KeyError: "not all values found in index 'lat'"

I would expect

>>> ds.sel(lat=[5,15,40], method="nearest", tolerance=5)
<xarray.DataArray (lat: 2)>
array([1, 2])
Coordinates:
  * lat      (lat) int64 10 20

Anything else we need to know?:

Environment:

Start at the DataArray.sel entry point and trace the nearest-indexing path used with tolerance. Reproduce the example as a regression test, then verify that out-of-tolerance labels are dropped while valid labels remain selected. Done means the failing example no longer raises KeyError and the expected two-item result is covered by tests.