django-deploy-probes

Production-ready deployment probes for Django applications, with HTTP endpoints and an in-process CLI runner.

Use django-deploy-probes for CI/CD deployment validation, Docker health checks, Kubernetes probes, blue-green deployments, and rollback checks.

Documentation

• Documentation site
• Basic setup
• Endpoint reference
• Security options
• Docker integration
• Docker Compose integration
• Kubernetes probes
• GitHub Actions deployment validation

Install

pip install django-deploy-probes

Supported runtimes:

• Python 3.10 through 3.14
• Django 5.2 LTS and Django 6.0

The 0.3.x release line remains available for projects that still require Django 4.2.

Optional extras:

pip install "django-deploy-probes[redis]"
pip install "django-deploy-probes[celery]"
pip install "django-deploy-probes[openapi]"
pip install "django-deploy-probes[all]"

Quick Start

INSTALLED_APPS = [
    "django_deploy_probes",
]

from django.urls import include, path

urlpatterns = [
    path("", include("django_deploy_probes.urls")),
]

Verify over HTTP:

curl -f http://localhost:8000/healthz
curl -f http://localhost:8000/readyz
curl -f http://localhost:8000/version

Or run the same probes in-process from Django:

python manage.py deploy_probes healthz --json
python manage.py deploy_probes readyz --json
python manage.py deploy_probes startupz --json
python manage.py deploy_probes version --json

The CLI reuses the same probe payload contract as the HTTP endpoints. Use it for CI/CD steps, pre-deploy validation, container bootstrap checks, and local debugging. For Kubernetes liveness, readiness, and startup probes, keep using HTTP endpoints as the default integration.

Custom check messages are hidden by default. If you enable EXPOSE_CHECK_MESSAGES=True, do not include secrets or sensitive values in those messages.

Security checks use REMOTE_ADDR by default. If probes are accessed through a trusted reverse proxy, configure TRUSTED_PROXY_NETWORKS and CLIENT_IP_HEADER to resolve the original client IP safely.

Storage checks use Django storage aliases, so the same probe can validate local filesystems, S3-compatible backends, and any custom storage backend wired through STORAGES.

Common settings

DEPLOY_PROBES = {
    "SERVICE_NAME": "my-django-app",
    "ENVIRONMENT": "prod",
    "VERSION": "1.2.0",
    "READY_CHECKS": ["database", "redis", "celery", "storage"],
    "STARTUP_CHECKS": ["migrations"],
    "READY_CUSTOM_CHECKS": [],
    "STARTUP_CUSTOM_CHECKS": [],
    "DATABASES": ["default"],
    "STORAGE": {
        "default": {
            "CHECK": "exists",
            "PATH": "probes/ready.txt",
        },
        "s3_media": {
            "CHECK": "write",
            "PREFIX": "deploy-probes",
        },
    },
    "REDIS": {
        "default": {
            "LOCATION": "redis://localhost:6379/0",
            "TIMEOUT": 1.0,
        },
    },
    "CELERY": {
        "BROKER": True,
        "WORKERS": False,
        "RESULT_BACKEND": False,
        "TIMEOUT": 1.0,
    },
    "DETAIL_LEVEL": "none",
    "INCLUDE_CHECK_DURATIONS": False,
    "REQUIRE_READY_CHECKS": False,
    "REQUIRE_STARTUP_CHECKS": False,
    "EXPOSE_CHECK_MESSAGES": False,
    "INTERNAL_IP_ONLY": False,
    "INTERNAL_IP_NETWORKS": [
        "127.0.0.1/32",
        "::1/128",
        "10.0.0.0/8",
        "172.16.0.0/12",
        "192.168.0.0/16",
    ],
    "TRUSTED_PROXY_NETWORKS": [],
    "CLIENT_IP_HEADER": None,
}

STORAGE supports two practical modes:

• CHECK="exists": verify that a known probe object exists.
• CHECK="write": create and delete a temporary object to verify write access.

For S3, exists is safer when you already manage a sentinel object like probes/ready.txt. Use write when you want to validate bucket write/delete permissions during readiness checks.

Timeout contract

There is no global probe timeout. Redis and Celery apply their TIMEOUT values only to operations whose client APIs support a real timeout:

• REDIS[alias]["TIMEOUT"]: Redis connect and socket timeout.
• CELERY["TIMEOUT"]: Celery broker connection and worker ping timeout.

Database, migration, storage, Celery result backend, and custom checks use their backend-native timeout configuration. Checks run sequentially, so set Kubernetes timeoutSeconds or an upstream load balancer timeout above the worst-case sum of the enabled checks.

With DETAIL_LEVEL="safe", recognized timeout exceptions return {"status": "fail", "reason": "timeout"}. See the timeout contract for configuration guidance.

Development

uv sync --dev
uv run pytest -q
uv run mkdocs build --strict

Publishing is handled by .github/workflows/publish.yml when a GitHub release is published. Documentation is deployed to GitHub Pages from main by .github/workflows/docs.yml.