add span reparenting field child_ids

[graphaelli]

Motivation/summary
Span reparenting intake. Closes #3422

Checklist

• I have signed the Contributor License Agreement.
• My code follows the style guidelines of this project (run make check-full for static code checks and linting)
• I have rebased my changes on top of the latest master branch

• I have made corresponding changes to the documentation
• I have added tests that prove my fix is effective or that my feature works
• New and existing unit tests pass locally with my changes

• I have updated CHANGELOG.asciidoc

How to test these changes

Related issues

• Relates Reparenting spans to support inferred spans apm#247

[graphaelli]

Even though it was in the proposal, I'm not sure about child_ids in the es document - child.id, which as keyword field would support >1 value, would be more symmetrical with the rest of the fields.

[axw]

+1 on child.id, but would like to hear @simitt's thoughts given her recent ECS research. Otherwise LGTM.

[graphaelli]

Good call. I pushed the change to child.id along with a changelog entry and will await feedback from @simitt.

[simitt]

IMO child is the right term here. What makes it confusing is that parent now doesn't actually mean parent anymore but can refer to any predecessor. A child's event can have another parent, that breaks my parent/child logic. Since parent.id is used for the waterfall UI (afaik) I don't think we could easily change that though, so LGTM.