Ingest node: terminate processor

[andrewkroh]

Description

Add a new ingest processor to terminate processing. When the processor is invoked, then execution of the ingest pipeline stops (control returns to calling pipeline if there is one).

A pattern we use today in pipelines to abort additional processing is to invoke the fail processor which jumps to the pipeline's on_failure handler, but this has the unwanted side-effect of being associated with a failure state and invokes additional processors (on_failure).

This pipeline is an example of where this would be used. I would rather abort the processing at this point without adding a failure message.

[elasticsearchmachine]

Pinging @elastic/es-data-management (Team:Data Management)

[PeteGillinElastic]

I didn't realize that I also need to edit processors.asciidoc to actually make the docs show up. PR for that out now.

[andrewkroh]

Is updating elasticsearch-specification normally in-scope for when a new processor is added? This would enable derivatives of the spec, like API clients and the Kibana Dev Console autocomplete, to know about the processor.

https://github.com/elastic/elasticsearch-specification/blob/53edb7836cd4d15fc15c1341a5eb7bf9dfb5d50a/specification/ingest/_types/Processors.ts#L31

[dakrone]

@andrewkroh it's supposed to be in scope whenever we make user-facing changes or additions, but we frequently forget about it. Thanks for the reminder!