feat: decode prom requests to grpc

[v0y4g3r]

I hereby agree to the terms of the GreptimeDB CLA.

Refer to a related PR or issue link (optional)

A follow-up in #3478 to improve ~40% more.

What's changed and what's your intention?

This PR decodes WriteRequest in prometheus protocol directly to GreptimeDB's RowInsertRequest to avoid the overhead of conversions.

The benchmark results show that it can reduce 32% deserialization time, hence a 47% improvement in throughput.

test result: ok. 0 passed; 0 failed; 2 ignored; 0 measured; 57 filtered out; finished in 0.00s

     Running benches/bench_prom.rs (target/release/deps/bench_prom-f03f7c5a882a859e)
decode/write_request    time:   [1.7256 ms 1.7287 ms 1.7322 ms]
                        change: [-87.783% -87.722% -87.657%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 2 outliers among 100 measurements (2.00%)
  1 (1.00%) high mild
  1 (1.00%) high severe
Benchmarking decode/prom_write_request: Warming up for 3.0000 s
Warning: Unable to complete 100 samples in 3.0s. You may wish to increase target time to 6.0s, enable flat sampling, or reduce sample count to 40.
decode/prom_write_request
                        time:   [1.1701 ms 1.1709 ms 1.1717 ms]
                        change: [-86.146% -86.130% -86.114%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 1 outliers among 100 measurements (1.00%)
  1 (1.00%) high mild

     Running benches/prom_decode.rs (target/release/deps/prom_decode-6fbb6196f7920b2a)

running 0 tests

This PR does not involves object pooling since it does little to decode performance but brings complexity in object lifecycle management. Memory pooling may help in this case instead.

Checklist

• I have written the necessary rustdoc comments.
• I have added the necessary unit tests and integration tests.
• This PR does not require documentation updates.

[codecov]

Codecov Report

Attention: Patch coverage is 60.62874% with 263 lines in your changes are missing coverage. Please review.

Project coverage is 84.49%. Comparing base (ae2c18e) to head (8c0c55b).
Report is 3 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #3425      +/-   ##
==========================================
- Coverage   85.08%   84.49%   -0.59%     
==========================================
  Files         889      893       +4     
  Lines      146488   146726     +238     
==========================================
- Hits       124632   123978     -654     
- Misses      21856    22748     +892     

[evenyag]

Most LGTM

[evenyag]

For performance reasons, we expect requests from the remote write protocol to be valid protobuf bytes. The protobuf standard says a string must always contain UTF-8 encoded or 7-bit ASCII text. However, we should provide a strict mode to perform UTF-8 validation or add this validation back once the string validation doesn't hurt the throughput too much.

[killme2008]

Great work! LGTM