Drift tracks pattern changes over time, showing regressions, improvements, and overall codebase health trends.

Trends analysis helps you:

• Detect pattern regressions before they become problems
• Track improvements in code quality
• Monitor category-specific trends
• Make data-driven decisions about technical debt

# View trends (default: 7 days)
drift trends

# 30-day trends
drift trends --period 30d

# 90-day trends with details
drift trends --period 90d --verbose

drift trends [options]

📊 Pattern Trends

Overall: 📈 IMPROVING
Period: 2024-01-08 → 2024-01-15

──────────────────────────────────────────────────
  Regressions:   3
  Improvements:  12
  Stable:        832
──────────────────────────────────────────────────

📉 Regressions (3):

  Critical:
    • Error Handling Coverage (errors)
      Coverage dropped from 85% to 72% (-15%)
    • API Response Validation (api)
      3 new endpoints without validation (+300%)

  Warning:
    • Test Coverage (testing)
      Coverage decreased slightly (-5%)

Category Trends:

  ↑ api: improving
  ↑ auth: improving
  ↓ errors: declining
  → security: stable
  → testing: stable

Use --verbose to see 12 improvements
View full details in the dashboard:
  drift dashboard

drift_trends({
  period?: "7d" | "30d" | "90d",  // Time period (default: 7d)
  category?: string,              // Filter by category
  limit?: number,                 // Max trends to return (default: 20)
  severity?: "all" | "critical" | "warning"  // Filter by severity (default: all)
})

Returns:

{
  "overallTrend": "improving",
  "startDate": "2024-01-08",
  "endDate": "2024-01-15",
  "regressions": [
    {
      "patternName": "Error Handling Coverage",
      "category": "errors",
      "severity": "critical",
      "changePercent": -15,
      "details": "Coverage dropped from 85% to 72%"
    }
  ],
  "improvements": [
    {
      "patternName": "API Documentation",
      "category": "documentation",
      "severity": "info",
      "changePercent": 25,
      "details": "Documentation coverage increased"
    }
  ],
  "stable": 832,
  "categoryTrends": {
    "api": { "trend": "improving", "change": 12 },
    "errors": { "trend": "declining", "change": -8 }
  }
}

Changes are calculated by comparing:

• Pattern counts (more/fewer instances)
• Coverage percentages
• Outlier counts
• Confidence scores

Every scan creates a snapshot in .drift/history/snapshots/:

.drift/history/snapshots/
├── 2024-01-08T10-00-00.json
├── 2024-01-09T10-00-00.json
├── 2024-01-10T10-00-00.json
└── ...

Trends compare the current state to historical snapshots:

Current State (today)
       │
       ▼
┌─────────────────┐
│  Compare with   │
│  7/30/90 days   │
│  ago            │
└────────┬────────┘
         │
         ▼
┌─────────────────┐
│  Calculate      │
│  - Regressions  │
│  - Improvements │
│  - Stable       │
└────────┬────────┘
         │
         ▼
┌─────────────────┐
│  Categorize     │
│  by severity    │
└─────────────────┘

A regression is detected when:

• Pattern count decreases significantly
• Coverage percentage drops
• Outlier count increases
• Confidence score decreases

Review code quality changes during a sprint:

drift trends --period 14d --verbose

Check for regressions before release:

drift trends --period 7d
# Ensure no critical regressions

Monitor long-term trends:

drift trends --period 90d --verbose

Track specific areas:

# Via MCP
drift_trends({ category: "security", period: "30d" })

Each category gets its own trend:

Run scans regularly to build history:

# Daily scan (cron or CI)
0 9 * * * cd /project && drift scan

Focus on high-impact categories:

# Check security and errors weekly
drift trends --period 7d | grep -E "(security|errors)"

After major refactors, establish new baselines:

drift scan
drift status --detailed
# Document current state as baseline

Don't let regressions accumulate:

# Check for critical regressions
drift trends --period 7d
# If critical regressions found, address immediately

Add trend checks to your CI pipeline:

# .github/workflows/trends.yml
name: Weekly Trends Check

on:
  schedule:
    - cron: '0 9 * * 1'  # Every Monday at 9am

jobs:
  trends:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
      
      - name: Install Drift
        run: npm install -g driftdetect
      
      - name: Check Trends
        run: |
          drift trends --period 7d > trends.txt
          if grep -q "DECLINING" trends.txt; then
            echo "⚠️ Code quality declining"
            cat trends.txt
            exit 1
          fi
      
      - name: Post to Slack
        if: failure()
        run: |
          curl -X POST $SLACK_WEBHOOK -d @trends.txt

View trends visually in the dashboard:

drift dashboard

The dashboard shows:

• Trend graphs over time
• Category breakdowns
• Regression highlights
• Improvement celebrations

Not enough history data to show trends.
Run more scans over time to see pattern trends.

Solution: Run regular scans to build history:

drift scan
# Wait and scan again later
drift scan
# Now trends will have data

If you see regressions after refactoring:

1. Check if patterns were intentionally changed
2. Review the specific files affected
3. Consider if the "regression" is actually an improvement

If a category doesn't appear in trends:

• No patterns in that category
• No changes in that category during the period

• Dashboard — Visual trend monitoring
• Quality Gates — Enforce trend requirements
• CI Integration — Automate trend checks