Motivation/summary

• Use json-iterator for faster decoding of _bulk responses.
• Use fastjson for faster (and fewer allocations) encoding of events for _bulk requests, and for encoding _bulk request metadata

We currently still use reflection for both encoding and decoding. We should generate code for both the response decoding and for encoding events and requests.

benchstat -delta-test=ttest /tmp/old.txt /tmp/new.txt 
name             old time/op    new time/op    delta
ModelIndexer-12    2.55µs ± 5%    2.03µs ±12%  -20.50%  (p=0.036 n=3+3)

name             old alloc/op   new alloc/op   delta
ModelIndexer-12    1.53kB ± 5%    1.31kB ± 1%  -14.15%  (p=0.031 n=3+3)

name             old allocs/op  new allocs/op  delta
ModelIndexer-12      12.0 ± 0%      10.0 ± 0%     ~     (zero variance)

Checklist

- [ ] Update CHANGELOG.asciidoc
- [ ] Documentation has been updated

How to test these changes

Non-functional change.

Related issues

#6360