Multi-core Analysis Support for GitLab Advanced SAST

Background

GitLab Advanced SAST currently performs deep code analysis which provides highly accurate results but leads to increased scan duration. While the improved accuracy is valuable for customers by reducing false positives compared to basic SAST solutions, the longer scan times can be problematic, especially for large monorepos.

Problem Statement

The current single-threaded analysis approach doesn't take advantage of modern multi-core processors, resulting in longer than necessary scan durations.

Proposed Solution

Implement parallel processing capabilities by:

Splitting the main ruleset into multiple smaller rulesets
Running multiple scan processes simultaneously across available CPU cores
Aggregating results from parallel scans into a single report

Expected Benefits

Reduced overall scan duration
Better resource utilization of modern hardware
Maintained high accuracy of results
Improved developer experience with faster feedback cycles

Technical Considerations

Need to ensure thread-safe operations
Proper resource allocation and management
Maintaining result consistency across parallel scans
Efficient ruleset partitioning strategy