feat: convert dot keys to nested for json logs

[nityanandagohain]

goos: darwin
goarch: arm64
pkg: github.com/SigNoz/signoz-otel-collector/utils/flatten
cpu: Apple M3 Pro
BenchmarkConvertJSON/NotJSON-11         	18709833	        64.03 ns/op	      64 B/op	       3 allocs/op
BenchmarkConvertJSON/Small-11           	  811808	      1505 ns/op	    1824 B/op	      33 allocs/op
BenchmarkConvertJSON/Medium-11          	  176397	      6908 ns/op	    7855 B/op	     123 allocs/op
BenchmarkConvertJSON/Large-11           	   49009	     23244 ns/op	   26815 B/op	     363 allocs/op
BenchmarkConvertJSON/VeryLarge-11       	   23659	     50417 ns/op	   62628 B/op	     926 allocs/op
PASS
ok  	github.com/SigNoz/signoz-otel-collector/utils/flatten	8.542s

the performance seems to be very bad .

I also tried a concurrent version where different go routines traveses different keys, but it has more overhead and it was more slower.

goos: darwin
goarch: arm64
pkg: github.com/SigNoz/signoz-otel-collector/utils/flatten
cpu: Apple M3 Pro
BenchmarkLargeJSONComparison/LargeJSON/Hybrid-11  	   10268	     99065 ns/op	   67045 B/op	     961 allocs/op
BenchmarkLargeJSONComparison/LargeJSON/ForcedConcurrent-11         	   12386	     96850 ns/op	   66991 B/op	     961 allocs/op
BenchmarkLargeJSONComparison/LargeJSON/ForcedNonConcurrent-11      	   19263	     62176 ns/op	   57917 B/op	     876 allocs/op
BenchmarkLargeJSONComparison/VeryLargeJSON/Hybrid-11               	   23764	     50303 ns/op	   62632 B/op	     926 allocs/op
BenchmarkLargeJSONComparison/VeryLargeJSON/ForcedConcurrent-11     	   15360	     78215 ns/op	   74391 B/op	    1099 allocs/op
BenchmarkLargeJSONComparison/VeryLargeJSON/ForcedNonConcurrent-11  	   23336	     50773 ns/op	   60319 B/op	     925 allocs/op
PASS
ok  	github.com/SigNoz/signoz-otel-collector/utils/flatten	12.006s