feat: initial twcs impl

[v0y4g3r]

I hereby agree to the terms of the GreptimeDB CLA

What's changed and what's your intention?

This PR adds support for TWCS compaction strategy inspired by Cassandra.

In previous leveled time window compaction strategy, SST files are strictly splitted and aligned to time windows. This may cause read amplification especially when a SST file contains rows from multiple time windows, even the out-of-order data is sparse.

TWCS computes SST file time window by their max timestamps in that out-of-order time-series data is mostly history data, it only affects min timestamps.

After assigning files into time windows, we then traverse every window to ensure that:

• for actively writing window, we allow for multiple SST files to alleviate write amplification
• for other windows, we allow at most one file per window for better read performance.

TWCS can be enable with table options:

CREATE TABLE IF NOT EXISTS cpu_metrics (
    hostname STRING,
    environment STRING,
    usage_user DOUBLE,
    usage_system DOUBLE,
    usage_idle DOUBLE,
    ts TIMESTAMP DEFAULT CURRENT_TIMESTAMP,
    TIME INDEX(ts),
    PRIMARY KEY(hostname, environment)

) WITH ('compaction'='twcs', 'compaction.twcs.max_active_window_files'=12, 'compaction.twcs.max_inactive_window_files'=3);

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)

Fixes #1068

[codecov]

Codecov Report

Merging #1851 (3179778) into develop (b8e9229) will decrease coverage by 0.32%.
The diff coverage is 83.78%.

@@             Coverage Diff             @@
##           develop    #1851      +/-   ##
===========================================
- Coverage    86.48%   86.17%   -0.32%     
===========================================
  Files          595      596       +1     
  Lines        96640    97143     +503     
===========================================
+ Hits         83582    83710     +128     
- Misses       13058    13433     +375     

[waynexia]

🚀