Add benchmark test for open_datatree #9100
Description
What is your issue?
To prevent regressions in the future, we should add a benchmark test for opening a deeply nested data tree.
This is to follow up on #9014 @aladinor's performance improvements.
You can see here how we benchmark opening and loading a single netCDF file
xarray/asv_bench/benchmarks/dataset_io.py
Line 125 in 447e5a3