feat: support fetch ranges in concurrent

[QuenKar]

I hereby agree to the terms of the GreptimeDB CLA

What's changed and what's your intention?

support concurrently fetch page ranges.
now the policy is that if object store support blocking, use blocking read priorly,
if not, use concurrently read.

Benchmark

Env

Local PC: mac m2 16G RAM 512 SSD

Config

[storage]
type = S3
region = minipc-0
root = "test-concurrent-read"
bucket = "test-bucket-0"

Result

fetch parquet according to ranges in sequence use develop branch, and this branch is concurrent.

Query	Use Sequence Blocking Read（mean）	Use Concurrent Read（mean）
cpu-max-all-1	1089.73ms	446.90ms
cpu-max-all-8	6287.86ms	1792.10ms
double-groupby-5	4230.23ms	3121.35ms
double-groupby-1	5581.11ms	4187.89ms
double-groupby-all	7625.86ms	6191.14ms
groupby-orderby-limit	2432.57ms	2731.98ms
high-cpu-1	345.14ms	229.95ms
high-cpu-all	7983.64ms	8101.72ms
lastpoint	12468.17ms	12795.19ms
single-groupby-1-1-1	134.69ms	126.53ms
single-groupby-1-1-12	160.07ms	157.99ms
single-groupby-1-8-1	371.19ms	397.88ms
single-groupby-5-1-1	152.95ms	222.45ms
single-groupby-5-1-12	183.41ms	222.23ms
single-groupby-5-8-1	554.36ms	528.04ms

conclusion

It's still slightly improved until lastpoint query，
because the lastpoint query basically reads the whole table, and all the pages are cached in memory, and there is not much difference in the following queries.

Checklist

• I have written the necessary rustdoc comments.
• I have added the necessary unit tests and integration tests.

Refer to a related PR or issue link (optional)

Closes #2808

[codecov]

Codecov Report

Attention: 27 lines in your changes are missing coverage. Please review.

Comparison is base (6221e5b) 85.90% compared to head (a6265e4) 85.39%.
Report is 1 commits behind head on develop.

Additional details and impacted files

@@             Coverage Diff             @@
##           develop    #2959      +/-   ##
===========================================
- Coverage    85.90%   85.39%   -0.51%     
===========================================
  Files          780      780              
  Lines       125900   125980      +80     
===========================================
- Hits        108155   107585     -570     
- Misses       17745    18395     +650     

[killme2008]

A great job! But I think we can do a benchmark for this feature. We don't know if this feature brings the performance upgrading or downgrading.

[evenyag]

A great job! But I think we can do a benchmark for this feature. We don't know if this feature brings the performance upgrading or downgrading.

@QuenKar Could you please add the benchmark result to the description?

[evenyag]

LGTM

[WenyXu]

Similar idea: apache/opendal#3675