feat(storage): adjust base compaction limitation

[Li0k]

I hereby agree to the terms of the RisingWave Labs, Inc. Contributor License Agreement.

What's changed and what's your intention?

In some scenarios, I found that CN generates small ssts, some ssts will be trivial-moved to base level, and the small files may destroy the efficiency of base level compaction, I want to adjust the strategy to reject some trivial-move and base level compaction tasks for some small ssts, so as to achieve better batching and reduce the small sst.

Notes: however, that this PR is not intended to significantly increase compaction capacity, but rather is a safeguard to minimise the number of small files compacting to the base level.

This PR changed:

1. Reject to base trivial move tasks that are less than the threshold when validator is turned on
2. Decrease the priority of base level compaction tasks that are less than the threshold (unexpected)

Checklist

• I have written necessary rustdoc comments
• I have added necessary unit tests and integration tests
• I have added test labels as necessary. See details.
• I have added fuzzing tests or opened an issue to track them. (Optional, recommended for new SQL features Sqlsmith: Sql feature generation #7934).
• My PR contains breaking changes. (If it deprecates some features, please create a tracking issue to remove them in the future).
• All checks passed in ./risedev check (or alias, ./risedev c)
• My PR changes performance-critical code. (Please run macro/micro-benchmarks and show the results.)

• My PR contains critical fixes that are necessary to be merged into the latest release. (Please check out the details)

Documentation

• My PR needs documentation updates. (Please use the Release note section below to summarize the impact on users)

Release note

If this PR includes changes that directly affect users or other significant modifications relevant to the community, kindly draft a release note to provide a concise summary of these changes. Please prioritize highlighting the impact these changes will have on users.

[Li0k]

https://buildkite.com/risingwave-test/nexmark-benchmark/builds/4308#0191331c-145d-428c-b3b3-34e609daf871

[Li0k]

P50 Base Level Compaction Task Size Distribution

main

branch

With the nexmark test (NEXMARK_QUERY="q0,q4,q5,q5-rewrite,q7,q7-rewrite,q8,q9,q12,q13,q3-no-condition"), We can observe that branch no longer selects compaction tasks below 128m.

[hzxa21]

I guess the motivation to change && to || here is that we choose to rely on L0 intra compaction to reduce the level count when ret.sstable_infos.len() >= self.min_expected_level_count so that we can ensure that to base compaction must contain at least self.min_expected_level_count and at least self.min_compaction_bytes. Let's document this in the codes.

[Li0k]

Consider a scenario

1. the sst at the bottom of each partition is too small to trigger trivial-move
2. there are too many sub levels to trigger a write-stop, and each partition is insufficient for Intra
3. all partitions do not satisfy the to base min_bytes and cannot to base

[hzxa21]

LGTM

[Li0k]

nexmark test:

• https://buildkite.com/risingwave-test/nexmark-benchmark/builds/4396#0191745e-cc90-4f44-9252-596c2291015d

[Li0k]

• perf(compactor): Record changes related to the compactor component #15973