[Website]: Reroute processor clarification on what happens next with the document

[ivosh]

Type of issue

None

What documentation page is affected

https://www.elastic.co/docs/reference/enrich-processor/reroute-processor

What happened?

The documentation of reroute processor should be improved, describing what happens next with the document once it is re-routed.
The page currently well describes what happens within the current pipeline (the paragraph starting with "After a reroute processor has been executed..."). However it does not describe that the document is processed in an ingest pipeline which corresponds to the new destination (or dataset+namespace). I am pretty sure there is some logic there which is worth describing. Most importantly, it should be described what happens if the new pipeline also contains a reroute processor. Could the document be rerouted several times? What will happen if there is a routing cycle?

Additional info

Thank you for a comprehensive Elastic documentation. I find it very valuable. If you could improve this page, it would be beneficial for all of us.

[theletterf]

Thanks @ivosh! We'll look into this.

[leemthompo]

@felixbarny would be interested in your perspective here as the author of this feature.

[felixbarny]

@ivosh made great suggestions that are definitely worth documenting.

However it does not describe that the document is processed in an ingest pipeline which corresponds to the new destination (or dataset+namespace).

That's right. It invokes the pipeline of the new target.

it should be described what happens if the new pipeline also contains a reroute processor.

Then the document is rerouted to yet another target. This happens as long as there's a pipeline that doesn't reroute it again.

Could the document be rerouted several times?

Yes.

What will happen if there is a routing cycle?

We're detecting cycles and are throwing an exception in that case.

[leemthompo]

@felixbarny nice, if you wanted to throw up a PR at some point, I could help with editorial. The page could probably do with some headings for logical groups and scannability, for example.

[felixbarny]

Not sure when I'll get to it. I'll keep it in the back burner. Feel free to pick this up and I can provide guidance and reviews.

[ivosh]

@felixbarny Thank you for the explanation. I think we have enough material for one additional paragraph for which @leemthompo could create a PR.

Here is my take:
After the document is rerouted, it is processed in an ingest pipeline of the new target. If the new ingest pipeline contains a reroute processor, the document is rerouted again to yet another target. Routing cycles are detected and in such cases the document fails to be ingested.