server: decommission pre-checks should also include the range id of erroring ranges

When decommission pre-checks fail, they report the error that prevented (simulating) decommissioning replica replacement.

The error is useful to determine which constraints may be problematic and whether the user should add more nodes to a referenced constraint locality/attribute.

However, when the goal is not to add more nodes but rather decommission immediately -- meaning the operator will remove any zone configuration constraints blocking decommission, they need to resort to skipping pre-checks and letting the decommission stall and looking at logs.

Include the range ID(s) of blocking ranges in the decommission pre-check CLI output.

Jira issue: CRDB-42722

Epic CRDB-37617