Add memory Metric for Projections PK keys

[npakeer]

Changelog category (leave one):

• Improvement

Changelog entry (a user-readable short description of the changes that goes into CHANGELOG.md):

Metrics to Track memory used by Projections Primary Keys and Projections Index Granularity for all tables.

Documentation entry for user-facing changes

• Documentation is written (mandatory for new features)

[clickhouse-gh]

Workflow [PR], commit [99a5aa8]

Summary: ✅

---

AI Review

Summary

This PR extends system.asynchronous_metrics with four projection-specific memory metrics (TotalProjectionPrimaryKeyBytesInMemory, TotalProjectionPrimaryKeyBytesInMemoryAllocated, TotalProjectionIndexGranularityBytesInMemory, TotalProjectionIndexGranularityBytesInMemoryAllocated) and adds an integration test that validates increase/decrease behavior across inserts and TRUNCATE. I did not find new correctness, safety, concurrency, or performance issues in the current PR head, and existing bot-raised gaps appear addressed.

ClickHouse Rules

Item	Status	Notes
Deletion logging	➖
Serialization versioning	➖
Core-area scrutiny	✅
No test removal	✅
Experimental gate	➖
No magic constants	✅
Backward compatibility	✅
SettingsChangesHistory.cpp	➖
PR metadata quality	✅
Safe rollout	✅
Compilation time	✅
No large/binary files	✅

Final Verdict

• Status: ✅ Approve

[tavplubix]

btw do we already count marks and index granularity from projection parts?

[alexey-milovidov]

Do not merge before fixing "Logical error: '!res.empty()' (STID: 2508-3ea0)"

[npakeer]

btw do we already count marks and index granularity from projection parts?

@tavplubix From what I understood, MarkCache memory usage metric is current metric, and is automatically incremented when marks file for any part(including projection parts) is loaded, because the increment of the loaded bytes happens with in the Cache Class during the load.

IndexGranularity, and PK metrics are asynchronous metrics and they are calculated separately for each interval i.e "asynchronous_metrics_update_period_s " and these metrics for projections were missed. This PR calculates PK and index granularity for projections in separate metrics.

[clickhouse-gh]

LLVM Coverage Report

Metric	Baseline	Current	Δ
Lines	84.00%	84.00%	+0.00%
Functions	90.90%	90.90%	+0.00%
Branches	76.50%	76.50%	+0.00%

Changed lines: 100.00% (22/22) · Uncovered code

Full report · Diff report

[azat]

Instead of this we can use SYSTEM RELOAD ASYNCHRONOUS METRICS

[azat]

Also make sense to add index_granularity_bytes_in_memory and index_granularity_bytes_in_memory_allocated to system.projection_parts

[azat]

Do not merge before fixing "Logical error: '!res.empty()' (STID: 2508-3ea0)"

Has been fixed already, likely #102604

.

[npakeer]

Also make sense to add index_granularity_bytes_in_memory and index_granularity_bytes_in_memory_allocated to system.projection_parts

@azat Thanks for the review. I will work on adding these