Internal refactoring of some IngestService methods

[joegallo]

Internal cleanup/refactoring -- each commit is pretty much independent, I'm trying to shed some WIP from a bigger PR that I'm still bringing into existence, and these seemed like the best candidates to be merged sooner.

[elasticsearchmachine]

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

[masseyke]

I don't see any problems with it, but seems like a lot of effort just to move one argument up two spots doesn't it? I assume there's a reason for it?

[joegallo]

"you are in a maze of twisty little passages, all alike" <-- me

For as much time as I've spent in IngestService over the last few weeks, I keep stumbling on the order of the arguments to executeBulkRequest and going "wait, which handler is this 40 lines of code inline lambda again?". Reordering them gets bulkRequestModifier::markItemAsDropped above the fold (right after bulkRequestModifier::markItemAsFailed, rather than sorta dangling at the end there below the fold) and it's clearly the onDropped handler based on its name.

Also it gets order of the handler arguments into the same order between executeBulkRequest and executePipelines -- onDropped and onFailure come first, and let you signal that a document should be dropped or that it failed, and then onCompletion which is the last thing you should call when you've processed all the documents (so there's a "last-i-ness" to it both in terms of order of arguments and actual execution semantics).