kirby-cloudfront-invalidations

Kirby plugin for automatic slug invalidation in AWS CloudFront

Installation

Composer

composer require liftric/kirby-cloudfront-invalidations

Setup

If you rely on the default credential provider chain, you just need to set the distributionId:

'liftric.cloudfrontinvalidations.distributionID' => 'YOUR_CF_DISTRIBUTION_ID'

Options

liftric.cloudfrontinvalidations.distributionID

This must be set to your AWS CloudFront distribution ID (available in the AWS CloudFront Distributions dashboard).

liftric.cloudfrontinvalidations.awsAccessKeyID

(optional) This can be set to the AWS access key ID of your AWS account, if you are not using the default credential provider chain.

liftric.cloudfrontinvalidations.awsSecretAccessKey

(optional) This can be set to the AWS secret access key of your AWS account, if you are not using the default credential provider chain.

liftric.cloudfrontinvalidations.dependantUrlsForPage

(optional) This must be a function that returns what URL(s) should be cleared after a page modification.

liftric.cloudfrontinvalidations.dependantUrlsForFile

(optional) This must be a function that returns what URL(s) should be cleared after a file modification.

liftric.cloudfrontinvalidations.dependantUrlsForSite

(optional) This must be a function that returns what URL(s) should be cleared after a site modification.

Example config

An some more advanced config, we use a larger dependantUrlsForPage function to collect from the old page as well the new page all related pages, considers the search page and prepares the urls for both languages, de and en. Of course this example could be much cleaner and maybe a bit more generic in some circumstances but should only demonstrate a different use case.

In addition to that, we also exclude the /api route from caching by sending a custom header which disallows caching. Having done both, we are able to use AWS CloudFront for a larger customer without any problems.

<?php
return [
    'liftric.cloudfrontinvalidations.dependantUrlsForPage' => function ($hook, $page, $oldPage = null) {
        $pages = new Pages();
        foreach ([$page, $oldPage] as $p) {
            if (!$p) continue;
            $pages->add($p);
            $pages->add($p->parents());
            foreach ([$p->related_products(), $p->related_pages()] as $q) {
                if (!$q) continue;
                if ($q->isNotEmpty()) {
                    $pages->add($q->toPages(','));
                }
            }
        }
        $pages->add(new Page(['slug' => 'search']));
        $pages = $pages->toArray(function ($page) {
            return array($page->urlForLanguage('de'), $page->urlForLanguage('en'));
        });
        $pages = array_values($pages);
        $pages = array_merge(...$pages);
        return $pages;
    },
    'hooks' => [
        'route:after' => function ($result, $path) {
            if (Str::startsWith($path, 'api/')) {
                header('Cache-Control: private, no-cache, no-store');
            }
        }
    ]
];

License

MIT

Kudos

• Neil Daniels of The Streamable for creating the kirby-clear-cloudflare-cache plugin which basically served as a blueprint for this plugin.