[timggggggg]

Description

Expand benchmark coverage to compare transform with single modify/rename plugins, conditional modify chains, and transform-only workloads.

Add field-value mapping, tag extraction, log parsing, and field renaming across representative input sizes, with both homogeneous and reproducible mixed inputs.

Measure processing of already decoded events, including do_if evaluation, while excluding setup and event restoration. Verify equivalent complete results and preserve aligned terminal output, deterministic ordering, and percentage comparisons for ns/op, B/op, and allocs/op.

Command: make bench-transform.

Scenario	Variant	ns/op	B/op	allocs/op
assign_new5	transform	454.0	312	9
assign_new5	modify	150.4 (-66.9%)	0 (-100.0%)	0 (-100.0%)
trim4	transform	1462	2352	33
trim4	modify	103.0 (-93.0%)	0 (-100.0%)	0 (-100.0%)
rename/preserve/5	transform	2189	1352	43
rename/preserve/5	rename	370.8 (-83.1%)	0 (-100.0%)	0 (-100.0%)
logs/no_regex/mixed	transform	2751	3536	63
logs/no_regex/mixed	modify	2867 (+4.2%)	687 (-80.6%)	0 (-100.0%)
logs/regex/mixed	transform	15322	1864	35
logs/regex/mixed	modify	84371 (+450.7%)	2235 (+19.9%)	12 (-65.7%)
chains/mapping/16/mixed	transform	602.8	744	15
chains/mapping/16/mixed	modify_x16	301.3 (-50.0%)	0 (-100.0%)	0 (-100.0%)
chains/tags/16/mixed	transform	3711	3724	110
chains/tags/16/mixed	modify_x16	486.4 (-86.9%)	0 (-100.0%)	0 (-100.0%)
dynamic/trim_slice/mixed	transform	2829	5774	27
collections/filter_sum/64	transform	37339	50912	556

Fixes #1026 #1033