perf: improve metrics signature #8879
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Contributor
PR Reviewer Guide 🔍Here are some key observations to aid the review process:
|
Contributor
PR Code Suggestions ✨Explore these optional code suggestions:
|
Contributor
There was a problem hiding this comment.
Greptile Overview
Greptile Summary
This PR optimizes the performance-critical signature_without_labels function by replacing std::hash::DefaultHasher with ahash::AHasher, addressing a 14% performance bottleneck in histogram queries processing 450,000 series.
Key Changes
- Performance Optimization: Switched from
DefaultHashertoAHasherinsignature_without_labels()function (src/service/promql/value.rs:793) - Test Improvements: Updated tests to verify deterministic behavior instead of relying on hardcoded hash values, making tests more robust across hash algorithm changes
- Profiling Support: Added
split-debuginfo = "packed"to profliing profile for better debugging capabilities - Code Formatting: Minor formatting improvements to dependency lists in Cargo.toml
Performance Impact
The signature function is heavily used throughout the PromQL engine:
- Called in aggregation operations (sum, avg, count, etc.) to group metrics by label sets
- Used in binary vector operations for matching time series
- Critical for duplicate label set detection
AHasher from the ahash crate provides significantly faster hashing than DefaultHasher, especially for small keys like metric labels, while maintaining good hash distribution.
Confidence Score: 5/5
- This PR is safe to merge with minimal risk
- This is a well-isolated performance optimization with proper test coverage. The change from DefaultHasher to AHasher is a straightforward swap that maintains the same interface. The ahash dependency was already in use in the codebase, and tests were updated to verify deterministic behavior rather than specific hash values, making them more resilient to future changes.
- No files require special attention
Important Files Changed
File Analysis
| Filename | Score | Overview |
|---|---|---|
| src/service/promql/value.rs | 5/5 | Replaced std::hash::DefaultHasher with ahash::AHasher in signature_without_labels function for better performance and updated tests to verify deterministic behavior instead of hardcoded hash values |
| Cargo.toml | 5/5 | Added profiling configuration (split-debuginfo = "packed") and reformatted dependencies for consistency |
Sequence Diagram
sequenceDiagram
participant Query as PromQL Query
participant Agg as Aggregation Module
participant Sig as signature_without_labels()
participant Hash as AHasher
participant Map as HashMap<u64, T>
Query->>Agg: Execute aggregation (e.g., sum, avg)
loop For each metric series
Agg->>Sig: Calculate signature(labels)
Sig->>Hash: Create AHasher::default()
Note over Hash: Previously: DefaultHasher::new()
loop For each label
Sig->>Hash: write(name.as_bytes())
Sig->>Hash: write(value.as_bytes())
end
Sig->>Hash: finish()
Hash-->>Sig: Return u64 hash
Sig-->>Agg: Return signature
Agg->>Map: Group by signature
end
Map-->>Query: Aggregated results
2 files reviewed, no comments
Contributor
|
This is my benchmark result: https://github.com/hengfeiyang/hello_world_rust/tree/main/hash |
|
| Status | Total | Passed | Failed | Skipped | Flaky | Pass Rate | Duration |
|---|---|---|---|---|---|---|---|
| All tests passed | 365 | 343 | 0 | 19 | 3 | 94% | 4m 39s |
|
| Status | Total | Passed | Failed | Skipped | Flaky | Pass Rate | Duration |
|---|---|---|---|---|---|---|---|
| All tests passed | 365 | 343 | 0 | 19 | 3 | 94% | 4m 39s |
|
| Status | Total | Passed | Failed | Skipped | Flaky | Pass Rate | Duration |
|---|---|---|---|---|---|---|---|
| All tests passed | 365 | 341 | 0 | 19 | 5 | 93% | 5m 1s |
|
| Status | Total | Passed | Failed | Skipped | Flaky | Pass Rate | Duration |
|---|---|---|---|---|---|---|---|
| All tests passed | 365 | 345 | 0 | 19 | 1 | 95% | 4m 40s |
|
| Status | Total | Passed | Failed | Skipped | Flaky | Pass Rate | Duration |
|---|---|---|---|---|---|---|---|
| All tests passed | 365 | 342 | 0 | 19 | 4 | 94% | 4m 38s |
hengfeiyang
approved these changes
Oct 24, 2025
|
| Status | Total | Passed | Failed | Skipped | Flaky | Pass Rate | Duration |
|---|---|---|---|---|---|---|---|
| All tests passed | 365 | 342 | 0 | 19 | 4 | 94% | 4m 42s |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
the signature is take 14% percent of time for calculate 45w series histogram query.
split-debuginfo = "packed"to generate symbol table for profiling in macos