Skip to content

PERFORMANCE-OPTIMIZATION.md

Latest commit

 

History

History
116 lines (77 loc) · 4.54 KB

PERFORMANCE-OPTIMIZATION.md

File metadata and controls

116 lines (77 loc) · 4.54 KB

Performance Optimization

This document describes the performance optimizations added to both the frontend and backend of Auto Code.

Overview

Performance improvements target four areas:

  1. Async event loop (uvloop) -- Backend async operations run faster on Linux/macOS.
  2. Zustand useShallow selectors -- Frontend re-renders are minimized by using shallow equality checks on store selectors.
  3. Bundle visualization -- A build-time analysis tool identifies large dependencies.
  4. Database connection pooling -- The web backend uses configurable connection pool parameters.

Backend: uvloop Integration

uvloop is a drop-in replacement for Python's default asyncio event loop. It is based on libuv and provides 2-4x throughput improvement for async I/O operations.

Location: apps/backend/core/client.py (top-level initialization)

Behavior:

  • On Linux and macOS, uvloop.install() is called at module load time.
  • On Windows, the default proactor event loop is used (uvloop does not support Windows).
  • If uvloop is not installed, the code falls back gracefully to the standard event loop.

Installation:

pip install uvloop  # Linux/macOS only

No configuration is needed. The optimization is automatic when the package is present.

Frontend: Zustand useShallow Selectors

Zustand v4+ uses reference equality to decide whether a component re-renders. When a selector returns an object or array, every state change triggers a re-render even if the selected slice did not change. Wrapping selectors with useShallow from zustand/react/shallow switches to shallow equality, preventing unnecessary re-renders.

Pattern:

import { useShallow } from 'zustand/react/shallow';

// Before (re-renders on every state change)
const projects = useProjectStore((state) => state.projects);

// After (re-renders only when projects actually change)
const projects = useProjectStore(useShallow((state) => state.projects));

Scope: Applied to components that consume array or object state from stores, including KanbanBoard, Sidebar, TerminalGrid, GitHubIssues, GitLabIssues, TaskCreationWizard, Worktrees, and ProductivityDashboard.

A full audit of all 22 Zustand stores is documented in ZUSTAND_SELECTOR_AUDIT.md at the repository root.

Frontend: Bundle Visualization

The rollup-plugin-visualizer plugin generates an interactive HTML report of the renderer bundle composition after each build.

Location: apps/frontend/electron.vite.config.ts (renderer plugins)

Output: apps/frontend/out/renderer/stats.html

The report shows module sizes (raw, gzip, brotli) in a treemap. Use it to identify large dependencies and code-splitting opportunities.

Configuration:

visualizer({
  filename: './out/renderer/stats.html',
  open: !process.env.CI,  // auto-open locally, skip in CI
  gzipSize: true,
  brotliSize: true,
})

Web Backend: Connection Pool Parameterization

The web backend (apps/web-backend/) uses SQLAlchemy with configurable connection pool settings. SQLite connections use check_same_thread=False for thread safety; non-SQLite databases use pool parameters from environment variables.

Location: apps/web-backend/core/database.py

Environment variables (non-SQLite databases only):

Variable Default Description
DB_POOL_SIZE 10 Maximum connections in the pool
DB_MAX_OVERFLOW 20 Extra connections beyond pool_size
DB_POOL_TIMEOUT 30 Seconds to wait for a connection
DB_POOL_RECYCLE 3600 Seconds before recycling a connection
DB_ECHO false Log SQL queries

Running Performance Tests

Performance regression tests live in tests/performance/.

# Run all performance tests
pytest tests/performance/ -v

# Run a specific suite
pytest tests/performance/test_async_performance.py -v
pytest tests/performance/test_memory_performance.py -v
pytest tests/performance/test_frontend_performance.py -v
pytest tests/performance/test_backend_performance.py -v

# Run only benchmark-marked tests
pytest tests/performance/ -m benchmark -v

Test suites:

File What it tests
test_async_performance.py Async task creation, concurrency, event loop overhead, uvloop integration
test_backend_performance.py Backend operation benchmarks
test_frontend_performance.py Bundle size constraints, Zustand optimization patterns, build artifacts
test_memory_performance.py Memory monitor performance, pressure detection, cleanup verification, data structure efficiency