fastapi-lens

API usage intelligence for FastAPI — Lightweight middleware that identifies thriving, cold, and dead endpoints. Zero-infrastructure, high-performance monitoring.

What's New in v0.3.0 (The Maintenance Update)

The v0.3.0 release marks a milestone in stability and automation:

• Dynamic TTL Cleanup: Automatically purge old logs to keep your SQLite database lean and fast.
• Advanced Latency (Percentiles): Support for P50, P95, and P99 metrics to find real bottlenecks.
• Error Segregation: Distinct tracking for 4xx (Client) vs 5xx (Server) errors.
• Trusted Publishing (OIDC): Fully automated and secure deployments to PyPI via GitHub Actions.
• Concurrency Fixes: Internal asyncio.Lock mechanisms to prevent race conditions during startup.

📊 HTML Dashboard

Visualizing your API health has never been easier. Access a built-in, glassmorphism-style dashboard at /lens/dashboard.

Key Dashboard Features:

• Real-time Metrics: Traffic totals, Success rates, and Active route counts.
• Latency Distribution: Watch your P95 and P99 response times per route.
• Error Visualization: Red/Green health bars indicating error rates.
• Status Filtering: Quickly find "Dead" or "Cold" endpoints to clean up your codebase.

Installation

pip install fastapi-lens

🛠️ Step-by-Step Setup (The "Senior" Way)

The recommended way to use fastapi-lens is via the setup() method, which handles both middleware and router registration automatically.

1. Basic Integration

from fastapi import FastAPI
from fastapi_lens import LensMiddleware, LensConfig

app = FastAPI()

# One-liner setup
LensMiddleware.setup(
    app, 
    config=LensConfig(
        db_path="lens.db",
        ttl_days=14,           # Auto-delete records older than 14 days
        security_enabled=True,
        report_key="your-secret-passphrase"
    )
)

2. Access your data

• UI: http://localhost:8000/lens/dashboard?report_key=your-secret-passphrase
• JSON API: Send request with header X-Lens-Key: your-secret-passphrase to /lens/report.

NOTE: If you set security_enabled=False, you can access the dashboard and API without a report key, using the default URL http://localhost:8000/lens/dashboard and http://localhost:8000/lens/report.

The report looks like this:

{
  "generated_at": 1775243992.65774,
  "filters": {
    "days": null,
    "status": null
  },
  "summary": {
    "total_endpoints": 3,
    "total_requests": 71,
    "active": 3,
    "cold": 0,
    "dead": 0,
    "never_called": 0
  },
  "endpoints": [
    {
      "path": "/items",
      "method": "GET",
      "status": "active",
      "total_calls": 26,
      "error_4xx_count": 0,
      "error_5xx_count": 0,
      "error_rate_pct": 0,
      "success_rate_pct": 100,
      "avg_duration_ms": 272.79,
      "p50_duration_ms": 245.91,
      "p95_duration_ms": 446.79,
      "p99_duration_ms": 469.53,
      "max_duration_ms": 485.7,
      "last_called_at": 1775241265.46036,
      "first_called_at": 1775165996.83268,
      "days_since_last_call": 0
    }
  ]
}

As we said, we can filter by status: /lens/report?status={active | cold | dead | never_called} or by days: /lens/report?days=7

We also have /lens/report/top to get the top endpoints by calls.

{
  "generated_at": 1775244158.6899,
  "window_days": 7,
  "endpoints": [
    {
      "path": "/items",
      "method": "GET",
      "status": "active",
      "total_calls": 26,
      "error_4xx_count": 0,
      "error_5xx_count": 0,
      "error_rate_pct": 0,
      "success_rate_pct": 100,
      "avg_duration_ms": 272.79,
      "p50_duration_ms": 245.91,
      "p95_duration_ms": 446.79,
      "p99_duration_ms": 469.53,
      "max_duration_ms": 485.7,
      "last_called_at": 1775241265.46036,
      "first_called_at": 1775165996.83268,
      "days_since_last_call": 0
    },
    {
      "path": "/users/{user_id}",
      "method": "GET",
      "status": "active",
      "total_calls": 23,
      "error_4xx_count": 2,
      "error_5xx_count": 8,
      "error_rate_pct": 43.48,
      "success_rate_pct": 56.52,
      "avg_duration_ms": 86.2,
      "p50_duration_ms": 66.89,
      "p95_duration_ms": 167.1,
      "p99_duration_ms": 171.65,
      "max_duration_ms": 184.45,
      "last_called_at": 1775241273.30783,
      "first_called_at": 1775165996.89715,
      "days_since_last_call": 0
    },
    {
      "path": "/",
      "method": "GET",
      "status": "active",
      "total_calls": 22,
      "error_4xx_count": 0,
      "error_5xx_count": 0,
      "error_rate_pct": 0,
      "success_rate_pct": 100,
      "avg_duration_ms": 0.8,
      "p50_duration_ms": 0.62,
      "p95_duration_ms": 1.22,
      "p99_duration_ms": 2.22,
      "max_duration_ms": 2.69,
      "last_called_at": 1775241252.63773,
      "first_called_at": 1775165996.62768,
      "days_since_last_call": 0
    }
  ]
}

And finally, we also have: /lens/report/dead to get the dead endpoints.

{
  "generated_at": 1775244273.47241,
  "dead_endpoint_count": 0,
  "endpoints": [
    {
      "path": "/",
      "method": "GET",
      "status": "active",
      "total_calls": 22,
      "error_4xx_count": 0,
      "error_5xx_count": 0,
      "error_rate_pct": 0,
      "success_rate_pct": 100,
      "avg_duration_ms": 0.8,
      "p50_duration_ms": 0.62,
      "p95_duration_ms": 1.22,
      "p99_duration_ms": 2.22,
      "max_duration_ms": 2.69,
      "last_called_at": 1775241252.63773,
      "first_called_at": 1775165996.62768,
      "days_since_last_call": 0
    }
  ]
}

⚙️ Advanced Configuration

Architecture

fastapi-lens is designed to be non-blocking. Your request path is never slowed down.

Your FastAPI app │ ├── LensMiddleware (Starlette BaseHTTPMiddleware) │ ├── Intercepts every request (non-blocking) │ ├── Resolves route template (/users/42 → /users/{user_id}) │ ├── Pushes to asyncio.Queue (never slows the request) │ └── Background task flushes queue → SQLiteStorage (WAL mode) │ └── /lens/dashboard (The Premium UI) └── /lens/report (JSON APIRouter)

1. Intercept: Middleware records timestamps using time.perf_counter().
2. Queue: Data is pushed to an asyncio.Queue (O(1) operation).
3. Batch: A background task drains the queue and writes to SQLite in batches using WAL mode for high concurrency.
4. Cleanup: A 24-hour cycle task runs DELETE and VACUUM based on your ttl_days.

Contributing 🤝

Contributions are what make the open-source community an amazing place to learn, inspire, and create. Any contributions you make are greatly appreciated.

1. Fork the Project
2. Create your Feature Branch (git checkout -b feature/AmazingFeature)
3. Commit your Changes (git commit -m 'Add some AmazingFeature')
4. Push to the Branch (git push origin feature/AmazingFeature)
5. Open a Pull Request

Follow the steps in the CONTRIBUTING.md file.

📄 License

Distributed under the MIT License. See LICENSE for more information.