Document tcp and connect timeout paramters

Our connect timeout options are not properly documented, see https://github.com/dask/distributed/blob/main/distributed/distributed-schema.yaml#L752-L758
which simply states "No Comment" on https://docs.dask.org/en/stable/configuration.html#distributed-comm

A frequent misconfiguration of timeouts.connect can lead to clusters seemingly being stuck. We should make sure it is properly documented what the effect of these options is and provide a guideline when custom values are appropriate.