Add prw2 pool

[SungJin1212]

Followed up #6330. This PR adds a PRW2.0 proto and sync.pool to PRW2 to avoid lots of GC.
The #6330 PR needs to be merged first.

These are the results of the benchmark.

benchstat old.txt new.txt
goos: darwin
goarch: arm64
pkg: github.com/cortexproject/cortex/pkg/util/push
cpu: Apple M1 Max
                                     │   old.txt   │               new.txt                │
                                     │   sec/op    │    sec/op     vs base                │
_convertV2RequestToV1/100_series-10    50.97µ ± 1%   23.22µ ±  3%  -54.44% (p=0.000 n=10)
_convertV2RequestToV1/500_series-10    249.2µ ± 0%   111.4µ ±  3%  -55.28% (p=0.000 n=10)
_convertV2RequestToV1/1000_series-10   495.9µ ± 0%   225.4µ ± 14%  -54.56% (p=0.000 n=10)
geomean                                184.7µ        83.54µ        -54.76%

                                     │   old.txt    │               new.txt                │
                                     │     B/op     │     B/op      vs base                │
_convertV2RequestToV1/100_series-10    94.09Ki ± 0%   42.52Ki ± 0%  -54.80% (p=0.000 n=10)
_convertV2RequestToV1/500_series-10    466.7Ki ± 0%   208.9Ki ± 0%  -55.24% (p=0.000 n=10)
_convertV2RequestToV1/1000_series-10   931.8Ki ± 0%   416.2Ki ± 0%  -55.33% (p=0.000 n=10)
geomean                                344.6Ki        154.6Ki       -55.13%

                                     │   old.txt    │               new.txt               │
                                     │  allocs/op   │  allocs/op   vs base                │
_convertV2RequestToV1/100_series-10     1513.0 ± 0%    513.0 ± 0%  -66.09% (p=0.000 n=10)
_convertV2RequestToV1/500_series-10     7.515k ± 0%   2.515k ± 0%  -66.53% (p=0.000 n=10)
_convertV2RequestToV1/1000_series-10   15.016k ± 0%   5.016k ± 0%  -66.60% (p=0.000 n=10)
geomean                                 5.548k        1.864k       -66.41%

goos: darwin
goarch: arm64
pkg: github.com/cortexproject/cortex/pkg/cortexpb
cpu: Apple M1 Max
BenchmarkMarshallWriteRequestV2
BenchmarkMarshallWriteRequestV2/no-pool
BenchmarkMarshallWriteRequestV2/no-pool-10         	  135732	      8818 ns/op	   11152 B/op	     102 allocs/op
BenchmarkMarshallWriteRequestV2/byte_pool
BenchmarkMarshallWriteRequestV2/byte_pool-10       	  282535	      4106 ns/op	    3116 B/op	       2 allocs/op
BenchmarkMarshallWriteRequestV2/byte_and_write_pool
BenchmarkMarshallWriteRequestV2/byte_and_write_pool-10         	  293763	      3961 ns/op	    2772 B/op	       1 allocs/op
PASS

Which issue(s) this PR fixes:
Fixes #6324

Checklist

• Tests updated
• Documentation added
• CHANGELOG.md updated - the order of entries should be [CHANGE], [FEATURE], [ENHANCEMENT], [BUGFIX]