Caching and Apache

As you may or may not have noticed, our website is entirely comprised of static files. We do this despite the fact that our site has a large amount of somewhat dynamic content (for instance, the blog you're reading right now). This is accomplished by using a static site generator (Pelican) and letting Apache serve the output straight off the file system. The primary benefits include:

• Tracking changes and entire history of the site is as easy as keeping it under version control.
• No server side scripting or processing to speak of.
• We don't need to query a database.

These last two help cut down on the response time when a user requests content which is vitally important.

One side effect of taking this approach is that browsers can pretty aggressively cache the resources that compose your site. This can either be a really good or really bad thing depending on the nature of your site. Browsers caching your content can result in very quick render times, but users may be looking at stale content. As I mention earlier in the article, our site changes frequently as we update it with new posts and such. This means a user seeing stale content is a problem for us.

There are several ways you can tell a browser to avoid caching (or at least check for an updated version) primarily through the use of HTTP headers. The header we're going to be interested in is the cache-control header. A few others headers exist to manipulate caching behavior. If you want to dive into more information, take a look at this blog post.

We can use the Apache module mod_expires to manipulate this header.

Prior to enabling this module, you'll want to add rules to your virtual host file (you can also put this configuration information in .htaccess). Our rules allow clients to cache content for 1 month with the exception of html files. These expire immediately requiring the browser to check with our server before using a cached version of the content. Using this strategy, we get the benefits of caching where it matters (larger files like images and frequently requested files like CSS) while getting the guarantee users will always see fresh content.

<IfModule mod_expires.c>
    ExpiresActive on
    ExpiresDefault          "access plus 1 month"

    ExpiresByType text/html "access plus 0 seconds"
</IfModule>

Once you've modified your configuration, enable the mod_expires module.

sudo a2enmod expires

Then restart apache to ensure your changes take effect.

sudo service apache2 restart