5.6 KiB
title |
---|
Common Caddyfile Patterns |
Common Caddyfile Patterns
This page demonstrates a few complete and minimal Caddyfile configurations for common use cases. These can be helpful starting points for your own Caddyfile documents.
These are not drop-in solutions; you will have to customize your domain name, ports/sockets, directory paths, etc. They are intended to illustrate some of the most common configuration patterns.
- Static file server
- Reverse proxy
- PHP
- Redirect
www.
subdomain - Trailing slashes
- Wildcard certificates
- Single-page apps (SPAs)
Static file server
example.com {
root * /var/www
file_server
}
As usual, the first line is the site address. The root
directive specifies the path to the root of the site (the *
means to match all requests, so as to disambiguate from a path matcher)—change the path to your site if it isn't the current working directory. Finally, we enable the static file server.
Reverse proxy
Proxy all requests:
example.com {
reverse_proxy localhost:5000
}
Only proxy requests having a path starting with /api/
and serve static files for everything else:
example.com {
root * /var/www
reverse_proxy /api/* localhost:5000
file_server
}
PHP
With a PHP FastCGI service running, something like this works for most modern PHP apps:
example.com {
root * /srv/public
encode gzip
php_fastcgi localhost:9000
file_server
}
Customize the site root accordingly; this example assumes that your PHP app's webroot is within a public
directory—requests for files that exist on disk will be served with file_server
, and anything else will be routed to index.php
for handling by the PHP app.
The php_fastcgi
directive is actually just a shortcut for several pieces of configuration.
Redirect www.
subdomain
To add the www.
subdomain with an HTTP redirect:
example.com {
redir https://www.{host}{uri}
}
www.example.com {
}
To remove it:
www.example.com {
redir https://example.com{uri}
}
example.com {
}
To remove it for multiple domains at once; this uses the {labels.*}
placeholders which are the parts of the hostname, 0-indexed from the right (e.g. 0=com, 1=example-one, 2=www):
www.example-one.com, www.example-two.com {
redir https://{labels.1}.{labels.0}{uri}
}
example-one.com, example-two.com {
}
Trailing slashes
You will not usually need to configure this yourself; the file_server
directive will automatically add or remove trailing slashes from requests by way of HTTP redirects, depending on whether the requested resource is a directory or file, respectively.
However, if you need to, you can still enforce trailing slashes with your config. There are two ways to do it: internally or externally.
Internal enforcement
This uses the rewrite
directive. Caddy will rewrite the URI internally to add or remove the trailing slash:
example.com {
rewrite /add /add/
rewrite /remove/ /remove
}
Using a rewrite, requests with and without the trailing slash will be the same.
External enforcement
This uses the redir
directive. Caddy will ask the browser to change the URI to add or remove the trailing slash:
example.com {
redir /add /add/
redir /remove/ /remove
}
Using a redirect, the client will have to re-issue the request, enforcing a single acceptable URI for a resource.
Wildcard certificates
If you need to serve multiple subdomains with the same wildcard certificate, the best way to handle them is with a Caddyfile like this, making use of the handle
directive and host
matchers:
*.example.com {
tls {
dns <provider_name> [<params...>]
}
@foo host foo.example.com
handle @foo {
respond "Foo!"
}
@bar host bar.example.com
handle @bar {
respond "Bar!"
}
# Fallback for otherwise unhandled domains
handle {
abort
}
}
Note that you must enable the ACME DNS challenge to have Caddy automatically manage wildcard certificates.
Single-page apps (SPAs)
When a web page does its own routing, servers may receive lots of requests for pages that don't exist server-side, but which are renderable client-side as long as the singular index file is served instead. Web applications architected like this are known as SPAs, or single-page apps.
The main idea is to have the server "try files" to see if the requested file exists server-side, and if not, fall back to an index file where the client does the routing (usually with client-side JavaScript): try_files {path} /index.html
A typical SPA config usually looks something like this:
example.com {
root * /path/to/site
encode gzip
try_files {path} /index.html
file_server
}
If your SPA is coupled with an API or other server-side-only endpoints, you will want to use handle
blocks to treat them exclusively:
example.com {
encode gzip
handle /api/* {
reverse_proxy backend:8000
}
handle {
root * /path/to/site
try_files {path} /index.html
file_server
}
}