parser_json: use JSON.parse instead of .load

[Watson1978]

Which issue(s) this PR fixes:
Fixes #

What this PR does / why we need it:
Ref. #4813 (comment)
Fix wrong usage of JSON.load method.
We should use JSON.parse instead.

JSON.load performs unnecessary deserialisation.

irb(main):001> JSON.load('{ "json_class": "String", "raw": [72, 101, 108, 108, 111] }')
=> "Hello"
irb(main):002> JSON.parse('{ "json_class": "String", "raw": [72, 101, 108, 108, 111] }')
=> {"json_class" => "String", "raw" => [72, 101, 108, 108, 111]}

Docs Changes:

Release Note:

[daipom]

LGTM. Thanks!

[daipom]

I think backporting to LTS is unnecessary because no specific issues have been reported with this.

I cannot assume how much this could change Fluentd's behavior.
So, I'm a little concerned about backporting this to LTS.
I think we should not backport this unless we have a strong reason.

If anyone has an opinion that we should backport, please comment.