Add nprocs auto option to dask-worker CLI

[jacobtomlinson]

Partly solves #4373

I've moved nprocesses_nthreads from distributed.deploy.local into distributed.deploy.utils and then am making use of it in distributed.cli.dask_worker.

Setting nprocs to auto when starting a worker will now have the same behaviour as LocalCluster in terms of the number of processes and threads created.

dask-worker tcp://localhost:8786 --nprocs=auto

[quasiben]

Kicked CI

[jacobtomlinson]

Thanks @quasiben. Looks like it was my fault, new test filename wasn't unique.

[jrbourbeau]

Just merged the main branch to resolve some unrelated CI failures

[jrbourbeau]

Thanks @jacobtomlinson! This is a nice addition. Could you add a test that checks when using --nprocs auto the expected number of workers with the expected number of threads are created? Something similar to distributed/cli/tests/test_dask_worker.py::test_nprocs_negative should work here too

[jacobtomlinson]

Thanks for the feedback @jrbourbeau. Added a test.

[jrbourbeau]

Thanks @jacobtomlinson! This LGTM, will merge after CI finishes