vmod_headerproxy

Varnish Header Proxy VMOD

Author:	Steve Easley
Date:	2016-05-20
Version:	0.1.0
Manual section:	3

This vmod requires Varnish 4.1 or higher.

SYNOPSIS

import headerproxy;

sub vcl_recv {
    headerproxy.call(``BACKEND``, ``PATH``);
}

sub vcl_deliver {
    headerproxy.process();
}

DESCRIPTION

Varnish vmod that enables you to add custom HTTP headers to the client request and response via whatever native programming language you prefer (PHP, Node, Ruby, Python, etc).

This wizardry is made possible by the vmod passing a copy of every client request to a web script (URL) you host. This should be a super fast script which returns a json response containing the headers the vmod will add to the request and response.

What can you do with this imbued power?

• Do additional processing on every client request that would be hard to do in VCL alone. Examples include logging, analytics, data warehousing, etc.
• Move complex business logic for controlling caching decisions out of VCL and into your web script. Examples include varying responses on geo location, AB testing, etc.
• Move complex VCL logic into your native application domain, where you and your developers can contribute more easily.
• Remove the need for numerous 3rd party vmods.

Check out the Wiki for additional details.

Check out the canonical example for inspiration.

FUNCTIONS

call

Prototype

headerproxy.call(BACKEND backend, STRING path)

Context

vcl_recv

Returns

VOID

Description

Tells the vmod to send a copy of the client request to your web script, decodes its json response, then inserts any requested request headers into the client request.

Example

# If your web script is hosted with your backend app, you can use the
# same backend/director your app uses like this:
sub vcl_recv {
    headerproxy.call(req.backend_hint, "/webscript");
}

# Or alternatively send requests to a dedicated director just for your
# proxy script. This assumes you have already created proxy_cluster .
sub vcl_recv {
    headerproxy.call(proxy_cluster.backend(), "/webscript");
}

process

Prototype

headerproxy.process()

Context

vcl_deliver

Returns

VOID

Description

Tells the vmod to inserts the requested response headers from your json.

Example

sub vcl_deliver {
    headerproxy.process();
}

error

Prototype

headerproxy.error()

Context

vcl_recv

Returns

STRING

Description

Called after headerproxy.call(), headerproxy.error() will return any error that might have occurred (as a string). Errors include CURL errors and JSON decoding errors. It will be empty if there were no errors.

Example

sub vcl_recv {
    headerproxy.call();
    set req.http.X-VMOD-Error = headerproxy.error();
}

INSTALLATION

The source tree is based on autotools to configure the building, and does also have the necessary bits in place to do functional unit tests using the varnishtest tool.

Usage:

./autogen.sh
./configure

If you have installed Varnish to a non-standard directory, call autogen.sh and configure with PKG_CONFIG_PATH pointing to the appropriate path. For example, when varnishd configure was called with --prefix=$PREFIX, use

PKG_CONFIG_PATH=${PREFIX}/lib/pkgconfig export PKG_CONFIG_PATH

Make targets:

• make - builds the vmod
• make install - installs your vmod in VMODDIR
• make check - runs the unit tests in src/tests/*.vtc

DEBUGGING

Configure vmod for debugging with configure --enable-debug. Useful debugging data will be outputted to both the Varnish log.

LIMITATIONS

• SSL responses from the web script url are currently not supported.

COMMON PROBLEMS

• configure: error: Need varnish.m4

  Check if PKG_CONFIG_PATH has been set correctly before calling autogen.sh and configure.

• No package 'libcurl' found

  Make sure libcurl-devel is installed.