XecureCode Reliability SDK - Python

AI-driven reliability SDK for Python applications.

The official Python SDK for the XecureCode Reliability Platform — a human-first failure analysis and recovery recommendation system.

---

Why This SDK Exists

Modern backend systems fail in unpredictable ways.

This SDK:

• Captures structured runtime errors
• Generates deterministic fingerprints
• Classifies error types (DATABASE, NETWORK, VALIDATION, etc.)
• Determines severity
• Sends non-blocking telemetry to your backend
• Never crashes your application

AI assists. Humans decide. Nothing executes automatically.

---

Requirements

• Python 3.8+
• FastAPI, Flask, or Django (optional)

---

Installation

pip install xecurecode-reliability-sdk

Or install with framework support:

pip install xecurecode-reliability-sdk[fastapi,flask,django]

---

Quick Start

1️⃣ Initialize the SDK

from reliability import ReliabilityClient, ReliabilityConfig

config = ReliabilityConfig(
    api_key="your-api-key",
    service_id="your-service",
    mode="development"
)
client = ReliabilityClient(config)

Configuration Options:

Parameter	Required	Description
api_key	Yes	Your API key
service_id	Yes	Service identifier
mode	Yes	development or production
timeout	No	Request timeout (default: 5000ms)

---

2️⃣ Automatic Global Error Capture

The SDK automatically handles:

• Uncaught exceptions
• Thread exceptions

# This will be automatically captured
raise ValueError("Database timeout")

---

3️⃣ FastAPI Integration

from fastapi import FastAPI
from reliability import ReliabilityClient, ReliabilityConfig
from reliability.fastapi_integration import FastAPIMiddleware

config = ReliabilityConfig(...)
client = ReliabilityClient(config)

app = FastAPI()
app.add_middleware(FastAPIMiddleware, client=client)

---

4️⃣ Flask Integration

from flask import Flask
from reliability import ReliabilityClient, ReliabilityConfig
from reliability.flask_integration import init_flask

config = ReliabilityConfig(...)
client = ReliabilityClient(config)

app = Flask(__name__)
init_flask(app, client)

---

5️⃣ Django Integration

In settings.py:

RELIABILITY_API_KEY = "your-api-key"
RELIABILITY_SERVICE_ID = "your-service"
RELIABILITY_MODE = "development"

In settings.py MIDDLEWARE:

MIDDLEWARE = [
    ...
    'reliability.django_integration.ReliabilityMiddleware',
    ...
]

---

6️⃣ Manual Capture

try:
    risky_operation()
except Exception as e:
    client.capture(e)

# With request context
client.capture(e, request)

---

What Gets Sent

Example structured payload:

{
  "message": "Database connection failed",
  "name": "ValueError",
  "fingerprint": "a8c39c21...",
  "timestamp": 1700000000000,
  "service": "my-service",
  "environment": "development",
  "severity": "critical",
  "errorType": "DATABASE",
  "serviceContext": {
    "pid": 12345,
    "runtime": "python",
    "runtimeVersion": "3.11.0",
    "pythonVersion": "3.11.0",
    "platform": "Linux-5.10.0",
    "hostname": "server-01"
  },
  "requestContext": {
    "method": "GET",
    "url": "/api/users",
    "ip": "10.0.0.5"
  },
  "occurrenceCount": 1
}

---

Error Classification

Type	Example
DATABASE	Timeout, SQL errors, connection issues
NETWORK	HTTP failures, socket errors
VALIDATION	Invalid input, type errors
AUTH	Permission denied
RUNTIME	Standard Python errors

---

Severity Detection

• Critical → Database, timeout, connection errors
• Medium → Validation, recoverable issues

---

Design Guarantees

• Non-blocking HTTP calls
• Timeout protected (5s default)
• Retry logic (3 attempts)
• Deduplication (1-minute window)
• Rate limiting (max 100 concurrent)
• Never throws inside SDK
• Never crashes host application
• Works with FastAPI, Flask, Django

---

License

MIT