SEO optimizations with Cloudformation

Looking (again) at SEO metrics, I wanted to fix two misbehaviors of the website: compression and error pages.
Let’s get through the process:

HTTP compression

This has been an easy one. The SEO tool wanted the site to accept compression, so moving from requesting this (locahost:4000 is the local hexo server where the html rendering is immediately visible):

GET / HTTP/1.1
Host: localhost:4000
Accept-Encoding: gzip, deflate, br

and getting no matching compression to asking for this:

GET / HTTP/1.1
Host: marcoaguzzi.it
Accept-Encoding: gzip, deflate, br

and be answered

Content-Encoding: br

which is the confirmation that Brotli compression is enabled.

Cloudformation snippet

The configuration should be done in the cloudfront serving the site hosted on S3. Namely, in the DefaultCacheBehavior by adding Compress: True to the yaml. The other settings required in the cache (enable gzip and enable brotli) are already served by referencing the pre-made CachingOptimized CachePolicy. Here’s the snippet: for clarity, the CachePolicyId is referenced in a map,
and, in this case, its value is 658327ea-f89d-4fab-a63d-7e88639e58f6

"DefaultCacheBehavior": {
    "Compress": true,
    ...,
    "CachePolicyId": {
        "Fn::FindInMap": [
        "CacheMapping",
        "Global",
        "CachingOptimized"
    ]
    },
    ...
}

References

• https://docs.aws.amazon.com/cloudfront/latest/APIReference/API_DefaultCacheBehavior.html
• https://docs.aws.amazon.com/AmazonCloudFront/latest/DeveloperGuide/ServingCompressedFiles.html
• https://docs.aws.amazon.com/AmazonCloudFront/latest/DeveloperGuide/using-managed-cache-policies.html

Custom error pages

The plugin managing the pagination of the website renders links to page “0” and “last page + 1” (they’re hidden, though). Crawling the website and pointing to those links, cloudfront replies with a 403 because the pages are not mapped in the S3 bucket. Instead of exposing the 403 error with an awkward cloudfront xml, it could be better to serve a page styled as the rest of the website, and maybe returning HTTP 200 code, thus the human user will see a courtesy page, but the http client won’t know that it hit a wrong path.

This can be achieved by instructing cloudfront to serve custom error pages when a HTTP error is raised. The path of the custom error pages must be present in the served S3 bucket.

Edit the Cloudformation template

In the DistributionConfig property section of CloudfrontDistribution resource, adding a CustomErrorResponses array will do the job. In this case only the HTTP 403 error is fired for unknown paths (per the S3 bucket configuration, there should be no 404 error),
so one line in the array will be enough. The object is quite self-explanatory: ErrorCode is the error fired, ResponseCode is the error code to be returned to the client, and ResponsePagePath is the page to be shown:

"CloudFrontDistribution": {
    "Properties": {
        "DistributionConfig": {
            "CustomErrorResponses":[
                {
                    "ErrorCode" : 403,
                    "ResponseCode" : 200,
                    "ResponsePagePath" : "/errors/403.html"
                }
            ],
        }
    }
}

Hexo page

The error page has been created like any other page in the site that is accessible from the topbar menu:

hexo new page errors

and a folder errors with an index.md file is created.
Next, the index.md file has been renamed to 403.md in order for it to be rendered as 403.html (with the hexo generate command), as the ResponsePagePath value states. The rendered html can then be pushed on git and deployed as any other article.
For reference, the actual error page is here.

The full Gist

The full parent cloudformation template can be found in this gist:

Click to view cloudformation template on gist