django-health-check

This project checks for various conditions and provides reports when anomalous behavior is detected.

The following health checks are bundled with this project:

• cache
• database
• storage
• disk and memory utilization (via psutil)
• AWS S3 storage
• Celery task queue
• Celery ping
• RabbitMQ
• Migrations

Writing your own custom health checks is also very quick and easy.

We also like contributions, so don't be afraid to make a pull request.

Use Cases

The primary intended use case is to monitor conditions via HTTP(S), with responses available in HTML and JSON formats. When you get back a response that includes one or more problems, you can then decide the appropriate course of action, which could include generating notifications and/or automating the replacement of a failing node with a new one. If you are monitoring health in a high-availability environment with a load balancer that returns responses from multiple nodes, please note that certain checks (e.g., disk and memory usage) will return responses specific to the node selected by the load balancer.

Supported Versions

We officially only support the latest version of Python as well as the latest version of Django and the latest Django LTS version.

Installation

First, install the django-health-check package:

$ pip install django-health-check

Add the health checker to a URL you want to use:

    urlpatterns = [
        # ...
        path(r'ht/', include('health_check.urls')),
    ]

Add the health_check applications to your INSTALLED_APPS:

    INSTALLED_APPS = [
        # ...
        'health_check',                             # required
        'health_check.db',                          # stock Django health checkers
        'health_check.cache',
        'health_check.storage',
        'health_check.contrib.migrations',
        'health_check.contrib.celery',              # requires celery
        'health_check.contrib.celery_ping',         # requires celery
        'health_check.contrib.psutil',              # disk and memory utilization; requires psutil
        'health_check.contrib.s3boto3_storage',     # requires boto3 and S3BotoStorage backend
        'health_check.contrib.rabbitmq',            # requires RabbitMQ broker
        'health_check.contrib.redis',               # requires Redis broker
    ]

Note: If using boto 2.x.x use health_check.contrib.s3boto_storage

(Optional) If using the psutil app, you can configure disk and memory threshold settings; otherwise below defaults are assumed. If you want to disable one of these checks, set its value to None.

    HEALTH_CHECK = {
        'DISK_USAGE_MAX': 90,  # percent
        'MEMORY_MIN': 100,    # in MB
    }

To use Health Check Subsets, Specify a subset name and associate it with the relevant health check services to utilize Health Check Subsets.

    HEALTH_CHECK = {
        # .....
        "SUBSETS": {
            "startup-probe": ["MigrationsHealthCheck", "DatabaseBackend"],
            "liveness-probe": ["DatabaseBackend"],
            "<SUBSET_NAME>": ["<Health_Check_Service_Name>"]
        },
        # .....
    }

To only execute specific subset of health check

curl -X GET -H "Accept: application/json" http://www.example.com/ht/startup-probe/

If using the DB check, run migrations:

$ django-admin migrate

To use the RabbitMQ healthcheck, please make sure that there is a variable named BROKER_URL on django.conf.settings with the required format to connect to your rabbit server. For example:

    BROKER_URL = "amqp://myuser:mypassword@localhost:5672/myvhost"

To use the Redis healthcheck, please make sure that there is a variable named REDIS_URL on django.conf.settings with the required format to connect to your redis server. For example:

    REDIS_URL = "redis://localhost:6370"

The cache healthcheck tries to write and read a specific key within the cache backend. It can be customized by setting HEALTHCHECK_CACHE_KEY to another value:

    HEALTHCHECK_CACHE_KEY = "custom_healthcheck_key"

Setting up monitoring

You can use tools like Pingdom, StatusCake or other uptime robots to monitor service status. The /ht/ endpoint will respond with an HTTP 200 if all checks passed and with an HTTP 500 if any of the tests failed. Getting machine-readable JSON reports

If you want machine-readable status reports you can request the /ht/ endpoint with the Accept HTTP header set to application/json or pass format=json as a query parameter.

The backend will return a JSON response:

    $ curl -v -X GET -H "Accept: application/json" http://www.example.com/ht/

    > GET /ht/ HTTP/1.1
    > Host: www.example.com
    > Accept: application/json
    >
    < HTTP/1.1 200 OK
    < Content-Type: application/json

    {
        "CacheBackend": "working",
        "DatabaseBackend": "working",
        "S3BotoStorageHealthCheck": "working"
    }

    $ curl -v -X GET http://www.example.com/ht/?format=json

    > GET /ht/?format=json HTTP/1.1
    > Host: www.example.com
    >
    < HTTP/1.1 200 OK
    < Content-Type: application/json

    {
        "CacheBackend": "working",
        "DatabaseBackend": "working",
        "S3BotoStorageHealthCheck": "working"
    }

Writing a custom health check

Writing a health check is quick and easy:

    from health_check.backends import BaseHealthCheckBackend

    class MyHealthCheckBackend(BaseHealthCheckBackend):
        #: The status endpoints will respond with a 200 status code
        #: even if the check errors.
        critical_service = False

        def check_status(self):
            # The test code goes here.
            # You can use `self.add_error` or
            # raise a `HealthCheckException`,
            # similar to Django's form validation.
            pass

        def identifier(self):
            return self.__class__.__name__  # Display name on the endpoint.

After writing a custom checker, register it in your app configuration:

    from django.apps import AppConfig

    from health_check.plugins import plugin_dir

    class MyAppConfig(AppConfig):
        name = 'my_app'

        def ready(self):
            from .backends import MyHealthCheckBackend
            plugin_dir.register(MyHealthCheckBackend)

Make sure the application you write the checker into is registered in your INSTALLED_APPS.

Customizing output

You can customize HTML or JSON rendering by inheriting from MainView in health_check.views and customizing the template_name, get, render_to_response and render_to_response_json properties:

    # views.py
    from health_check.views import MainView

    class HealthCheckCustomView(MainView):
        template_name = 'myapp/health_check_dashboard.html'  # customize the used templates

        def get(self, request, *args, **kwargs):
            plugins = []
            status = 200 # needs to be filled status you need
            # ...
            if 'application/json' in request.META.get('HTTP_ACCEPT', ''):
                return self.render_to_response_json(plugins, status)
            return self.render_to_response(plugins, status)

        def render_to_response(self, plugins, status):       # customize HTML output
            return HttpResponse('COOL' if status == 200 else 'SWEATY', status=status)

        def render_to_response_json(self, plugins, status):  # customize JSON output
            return JsonResponse(
                {str(p.identifier()): 'COOL' if status == 200 else 'SWEATY' for p in plugins},
                status=status
            )

    # urls.py
    import views

    urlpatterns = [
        # ...
        path(r'ht/', views.HealthCheckCustomView.as_view(), name='health_check_custom'),
    ]

Django command

You can run the Django command health_check to perform your health checks via the command line, or periodically with a cron, as follow:

    django-admin health_check

This should yield the following output:

    DatabaseHealthCheck      ... working
    CustomHealthCheck        ... unavailable: Something went wrong!

Similar to the http version, a critical error will cause the command to quit with the exit code 1.

Other resources

• django-watchman is a package that does some of the same things in a slightly different way.