Context: #2569 (comment)

In an ideal world, the ProcessorChain processors would support the exact same config options as regular top-level nodes (named outputs, entity configuration, header_capture, etc).

For the initial implementation, it makes sense to have a separate minimal configuration.

At the very least, NodeUserConfig would need to support an optional name property. It isn't required for regular nodes because name (and node.id) is already implied by config structure. For SubProcessors (which appear in list format) the user can optionally specify an override of the default index-based naming (chain_name/0, chain_name/1).