Behind the scenes: How we host Ars Technica, part 1 (arstechnica.com)

Wordpress is a versatile platform, that's for sure.

As a long term fan of ars, have considered abandoning it, due to the repetitive autoplay video in articles recently.

Always the same video!

Don’t see how it could do anything but alienate viewers while increasing hosting costs.

For some reason I was expecting it to be more interesting than just being all-in on AWS services. Its fine and a reasonable approach but kind of a yawn in my opinion.

That's quite good actually thinking about it. It's not an overly complex piece infrastructure which is unusual but wholly welcome for similar HN submissions.

ECS/Fargate and ElasticSearch makes a lot of sense here considering Wordpress, and keeps their stack as well as local development simple.

The levels of abstraction are improving, but we're still far away from true utility-level convenience. Ideally, you'd be able to put your code rules in Github and be done with it. Details of database, file storage, caching, scaling, etc. would be something you didn't have to think about at all.

This is very well written, and explains clearly how each AWS service works.

I have been interested in how one would deploy Wordpress in a high availability setting, and this is just that!

Thanks Lee for the great article, looking forward to the next parts.

Surprised they're using Fargate here rather than an EC2-based cluster.

We use Fargate but we have very spiky traffic and have a very small team.

I would have expected a larger company with a relatively predictable traffic load to use an EC2 cluster to cut down on cost.