Vaban

A quick and easy way to control clusters of Varnish Cache hosts using a RESTful JSON API.

Vaban is built in Go for high performance, concurrency and simplicity. Every request and every ban spawns its own lightweight thread. It supports Varnish 3.0.3 + 4, Authentication, Pattern-based/VCL-based banning, health status, enable/disable backends, and more stuff to come.

Getting Started

Installing from packages

The easiest way to install Vaban is from packages.

• Currently enabled for Ubuntu 14.04/12.04 and Debian 7.
• Current packages available from packager.io

Installing from source

Dependencies

• Git
• Go 1.4+

Clone and Build locally:

git clone https://github.com/martensson/vaban.git
cd vaban
go build

Create a config.yml file and add all your services:

Put the file inside your application root, if installing from package: /opt/vaban/config.yml

---
service1:
  hosts:
    - "a.example.com:6082"
    - "b.example.com:6082"
    - "c.example.com:6082"
    - "d.example.com:6082"
    - "e.example.com:6082"
  secret: "1111-2222-3333-aaaa-bbbb-cccc"

service2:
  hosts:
    - "x.example.com:6082"
    - "y.example.com:6082"
    - "z.example.com:6082"
  secret: "1111-2222-3333-aaaa-bbbb-cccc"

Running Vaban

from source:

./vaban -p 4000 -f /path/to/config.yml

from packages:

vaban run web

or

vaban scale web=1
service vaban start
vaban logs

from Docker

If you already have Docker install its really use to compile and run with two commands:

docker build -t vaban .
docker run -p 4000:4000 vaban

Make sure that the varnish admin interface is available on your hosts, listening on 0.0.0.0:6082

API

A quick and easy way to control clusters of Varnish Cache hosts using a RESTful JSON API.

GET /v1/services

Get all groups:

• Response 200 (application/json)

    [
        "group1",
        "group2",
        "group3",
    ]
  

GET /v1/service/group1

Get all hosts in group:

• Response 200 (application/json)

    [
        "test01:6082"
    ]
  

GET /v1/service/group1/ping

Scan hosts to see if tcp port is open:

• Response 200 (application/json)

    {
        "test01:6082": {
            "Msg": "PONG 1431078011 1.0"
        }
    }
  

GET /v1/service/group1/health

Check health status of all backends:

• Response 200 (application/json)

    {
        "test01:6082": {
            "backend01(10.160.101.100,,80)": {
                "Refs": "1",
                "Admin": "probe",
                "Probe": "Healthy 4/4"
            }
        }
    }
  

GET /v1/service/group1/health/www01

Check health status of one backend:

• Response 200 (application/json)

    {
        "test01:6082": {
            "backend01(10.160.101.100,,80)": {
                "Refs": "1",
                "Admin": "probe",
                "Probe": "Healthy 3/4"
            }
        }
    }
  

POST /v1/service/group1/health/www01

force health status of one backend (can be healthy, sick or auto):

• Request (application/json)

    {"Set_health":"sick"}
  

• Response 200 (application/json)

    {
        "test01:6082": {
            "Msg": "updated with status 200 0"
        }
    }
  

POST /v1/service/group1/ban

To ban elements in your cache.

• Request Ban the root of your website (application/json)

    {"Pattern":"/"}
  

• Request Ban all css files (application/json)

    {"Pattern":".*css"}
  

• Request Ban based on VCL, in this case all objects matching a host-header. (application/json)

    {"Vcl":"req.http.Host == 'example.com'"}"}
  

• Response 200 (application/json)

    {
        "test01:6082": {
            "Msg": "ban status 200 0"
        }
    }