Skip to content

Latest commit

 

History

History
238 lines (198 loc) · 5.99 KB

README.md

File metadata and controls

238 lines (198 loc) · 5.99 KB

Fail2ban plugin for traefik

Build Status

This plugin is an implementation of a Fail2ban instance as a middleware plugin for Traefik.

Middleware

After installing the plugin, it can be configured through a Middleware, e.g.:

apiVersion: traefik.io/v1alpha1
kind: Middleware
metadata:
  name: fail2ban-test
spec:
  plugin:
    fail2ban:
      logLevel: DEBUG
      denylist:
        ip: 127.0.0.1
Add the middleware to an ingressroute
apiVersion: traefik.io/v1alpha1
kind: IngressRoute
metadata:
  name: simplecrd
  namespace: default
spec:
  entryPoints:
    - web
  routes:
  - match: Host(`fail2ban.localhost`)
    kind: Rule
    middlewares:
    - name: fail2ban-test
    services:
    ...

Configuration

Please note that the allowlist and denylist functionality described below can only be used concurrently with Fail2ban functionality (if you are looking for a way to allowlist or denylist IPs without using any of the Fail2ban logic, you might want to use a different plugin.)

Allowlist

You can allowlist some IP using this:

testData:
  allowlist:
    files:
      - "tests/test-ipfile.txt"
    ip:
      - "::1"
      - "127.0.0.1"

Where you can use some IP in an array of files or directly in the configuration.

If you have a single IP, this: ip: 127.0.0.1 should also work.

Denylist

Like allowlist, you can denylist some IP using this:

testData:
  denylist:
    files:
      - "tests/test-ipfile.txt"
    ip:
      - "::1"
      - "127.0.0.1"

Where you can use some IP in an array of files or directly in the configuration.

Please note that Fail2ban logs will only be visible when Traefik's log level is set to DEBUG.

Fail2ban

We plan to use all default fail2ban configuration but at this time only a few features are implemented:

testData:
  rules:
    urlregexps:
    - regexp: "/no"
      mode: block
    - regexp: "/yes"
      mode: allow
    bantime: "3h"
    findtime: "10m"
    maxretry: 4
    enabled: true
    statuscode: "400,401,403-499"

Where:

  • findtime: is the time slot used to count requests (if there is too many requests with the same ip in this slot of time, the ip goes into ban). You can use 'smart' strings: "4h", "2m", "1s", ...
  • bantime: correspond to the amount of time the IP is in Ban mode.
  • maxretry: number of request before Ban mode.
  • enabled: allow to enable or disable the plugin (must be set to true to enable the plugin).
  • urlregexp: a regexp list to block / allow requests with regexps on the url
  • statuscode: a comma separated list of status code (or range of status codes) to consider as a failed request.

URL Regexp

Urlregexp are used to defined witch part of your website will be either allowed, blocked or filtered :

  • allow : all requests where the url match the regexp will be forwarded to the backend without any check
  • block : all requests where the url match the regexp will be stopped
No definitions
testData:
  rules: {}

By default, fail2ban will be applied.

Multiple definition
testData:
  rules:
    urlregexps:
    - regexp: "/whoami"
      mode: allow
    - regexp: "/do-not-access"
      mode: block

In the case where you define multiple regexp on the same url, the order of process will be :

  1. Block
  2. Allow

In this example, all requests to /do-not-access will be denied and all requests to /whoami will be allowed without any fail2ban interaction.

Status code

When this configuration is set (i.e., statuscode is not empty), the plugin will wait for the request to be completed and check the status code of the response. If the status code is in the list of status codes, the request will be considered as a failed request.

Note that the request is considered completed when the response is back sent to the plugin, therefore, the request went through the middleware, traefik, to the backend, and back to the middleware.

Here is a little schema to explain the process
sequenceDiagram
    actor C as Client
    participant A as Middleware
    participant B as Backend
    C->>A: Request
    A->>B: Request
    B->>A: Response
    A->>A: Check status code
    critical [Check status code]
    option Invalid status code
        A--X C: Log error
    option valid status code
        A->>C: Log error
    end
Loading

Schema

First request, IP is added to the Pool, and the findtime timer is started:

A |------------->
  ↑

Second request, findtime is not yet finished thus the request is fine:

A |--x---------->
     ↑

Third request, maxretry is now full, this request is fine but the next wont.

A |--x--x------->
        ↑

Fourth request, too bad, now it's jail time, next request will go through after bantime:

A |--x--x--x---->
           ↓
B          |------------->

Fifth request, the IP is in Ban mode, nothing happen:

A |--x--x--x---->
B          |--x---------->
              ↑

Last request, the bantime is now over, another findtime is started:

A |--x--x--x---->            |------------->
                             ↑
B          |--x---------->

How to dev

$ docker compose up

Authors

Tom Moulard Clément David Martin Huvelle Alexandre Bossut-Lasry