Blacklist users and hosts in Django. Automatically blacklist rate-limited clients.
Django Blacklist allows you to block specific users and IP addresses/networks from accessing your application. Clients can be blocked manually from the admin interface, or automatically after exceeding a request rate limit. Each blacklist rule is applied for a specific duration. The blacklist is very scalable and is applied without noticeable overhead for large numbers of rules.
To install the package, run:
$ pip install django-blacklist
Add the blacklist
application to INSTALLED_APPS
:
INSTALLED_APPS = [
...
'blacklist'
]
Add the BlacklistMiddleware
middleware after AuthenticationMiddleware
:
MIDDLEWARE = [
...
'django.contrib.auth.middleware.AuthenticationMiddleware',
'blacklist.middleware.BlacklistMiddleware',
...
]
Apply the blacklist database migrations:
$ python manage.py migrate blacklist
You can manage the blacklist rules from the admin.
Changes take effect after a configurable time, or when the server is restarted.
A rule can target a user or an IP address.
You can also target IP networks (ranges) by specifying the optional prefixlen
field (number of network prefix bits).
Each rule has a specific duration. After that time interval passes, the rule expires.
When a request is blocked due to a matching rule:
- Status 400 (bad request) is returned.
- An error template is rendered. You can specify a custom one (see Settings below), or use the one for status 400.
- A message is logged
(warning from logger
blacklist.middleware
for custom templates, or error from loggerdjango.security
otherwise).
Expired rules are not automatically removed from the database.
They can be cleaned up with the included management command trim_blacklist
:
$ python manage.py trim_blacklist [-c <created_days>] [-e <expired_days>]
The options -c
and -e
specify the minimum ages of creation and expiry, respectively.
Clients can be blacklisted automatically, after exceeding a specified request rate limit. This feature requires django-ratelimit.
First, rate-limit a view by applying the @ratelimit
decorator. Make sure to set block=False
.
Then, blacklist rate-limited clients by adding the @blacklist_ratelimited
decorator. Specify the blacklist duration.
For example:
from datetime import timedelta
from django_ratelimit.decorators import ratelimit
from blacklist.ratelimit import blacklist_ratelimited
@ratelimit(key='user_or_ip', rate='50/m', block=False)
@blacklist_ratelimited(timedelta(minutes=30))
def index(request):
...
Automatic rules take effect immediately. If the request comes from an authenticated user, the rule will target that user. Otherwise, it will target their IP address.
@blacklist_ratelimited
accepts two arguments: (duration, block=True)
.
duration
can be atimedelta
object, or a tuple of two separate durations (for user-based and IP-based rules).block
specifies if the request should be blocked immediately, or passed to the view.
Automatic rules will have a comment that contains the ID of the request, which triggered the creation of the rule, and the "request line". The request ID is added only if available. Django does not generate request IDs. For that purpose, you can install django-log-request-id.
By default, the client IP address is taken from the REMOTE_ADDR
value of request.META
.
If your application server is behind one or more reverse proxies,
this will usually be the address of the nearest proxy, and not the actual client address.
To properly blacklist clients by IP address,
you can configure Django Blacklist to use addresses from another source (see Settings
below).
To actually obtain the proxied client addresses,
you can use django-ipware.
In this case, you can configure Django Blacklist to obtain client addresses from your function,
which in turn calls django-ipware
for the actual logic.
Alternatively, you can set REMOTE_ADDR
from the X-Forwarded-For
header in middleware,
installed before Django Blacklist.
However, keep in mind that this header can be forged to bypass the rate limits.
To counter that, you can use the last address in that header (which should be set by your trusted reverse proxy).
If you are behind two proxies, use the second to last address, and so on.
BLACKLIST_ENABLE
- whether blacklisted clients should be blocked, and rate-limited clients should be blacklisted; default:True
BLACKLIST_RELOAD_PERIOD
- how often to reload the blacklist, in seconds; default:60
BLACKLIST_RATELIMITED_ENABLE
- whether rate-limited clients should be automatically blacklisted; requiresBLACKLIST_ENABLE
; default:True
BLACKLIST_TEMPLATE
- name of a custom error template to render to blocked clients; its context will containrequest
andexception
; set toNone
to use the template for status 400; default:None
BLACKLIST_LOGGING_ENABLE
- whether blocked requests should be logged (honored only if a custom error template is configured); default:True
BLACKLIST_ADDRESS_SOURCE
- the source of client addresses; can be a key inrequest.META
, a callable that receives the request object, or the dotted string path to such a callable; default:'REMOTE_ADDR'