ddog-monitoring 0.6.0

Lightweight, framework-agnostic Python library for sending metrics to Datadog, SumoLogic, or Console. Lambda/Zappa-ready with live config support.

ddog-monitoring

A lightweight, framework-agnostic Python library for sending metrics to Datadog, SumoLogic, or a local Console output. Designed for use in Django, FastAPI, or any Python app — including AWS Lambda / Zappa deployments.

Installation

pip install ddog-monitoring

Quick Start

from monitoring.client import MonitoringFacade
from monitoring.config import MonitoringConfig

# Configure once at startup
MonitoringFacade.configure(MonitoringConfig(
    provider="datadog",
    datadog_api_key="your-api-key",
    datadog_app_key="your-app-key",
    datadog_hostname="your-host",
))

# Use anywhere
provider = MonitoringFacade.get_instance()
provider.send("api.request", value=1, metric_type="count", tags=["env:prod"])

Providers

Provider	Config keys required
console	(none) — prints metrics to stdout
datadog	datadog_api_key, datadog_app_key, datadog_hostname
sumologic	sumologic_url

Django + AWS Lambda / Zappa Usage

For Django apps, call configure() once in AppConfig.ready(). Pass a callable factory (lambda) instead of a static config so that django-constance value changes are picked up automatically on warm Lambda starts — without a redeploy.

# main_app/apps.py
from django.apps import AppConfig

class MainAppConfig(AppConfig):
    name = "main_app"

    def ready(self):
        from constance import config
        from monitoring.client import MonitoringFacade
        from monitoring.config import MonitoringConfig

        MonitoringFacade.configure(lambda: MonitoringConfig(
            provider=config.MONITORING_PROVIDER,
            datadog_api_key=config.DATADOG_API_KEY,
            datadog_app_key=config.DATADOG_APP_KEY,
            datadog_hostname=config.DATADOG_HOSTNAME,
            sumologic_url=config.SUMOLOGIC_URL,
        ))

Create a thin accessor module so views, management commands, and middleware never import MonitoringFacade directly:

# main_app/monitor.py
from monitoring.client import MonitoringFacade

def get_monitor():
    return MonitoringFacade.get_instance()

# views.py
from .monitor import get_monitor

def my_view(request):
    get_monitor().send("api.hit", value=1, metric_type="count", tags=["endpoint:my_view"])

How warm-start caching works

On every get_instance() call the library fetches the current config (via your factory) and compares it using Pydantic field equality. The provider is only recreated when config actually changes — otherwise the cached instance is reused immediately, keeping warm starts fast.

Decorators

Track function execution time and errors automatically:

from monitoring.decorators import track_metric

@track_metric("my_function")
def my_function():
    ...

License

MIT