Flipping The AWS Switch - Quest for the One Blog, Part 15

Tonight I’m going to start “flipping the switch” to host the blog entirely in AWS.

You, constant blog reader, will not notice anything.

Unless something goes horribly wrong. In that case, everything will explode and I’ll need to figure out how to roll back the changes gracefully. You might not even get to see this post explaining what happened.

When it’s done, you might notice the site loads about a bazillion times faster, as it will no longer be served from an antiquated shared hosting platform. There will also be a different endgameviable.com certificate.

The site that you or your feed reader programs load will become a teeny tiny custom Golang program that runs on AWS Elastic Beanstalk. That program will decide if you (or, let’s be honest, Google) are trying to load a very old link that needs to be redirected. If so, it will redirect to the new, updated link. Otherwise it will make a reverse proxy call to load the blog content from a static Hugo-built site, hosted with AWS Amplify.

If you’re one of those people who runs 50,000 script and ad blockers, it’s possible they might mention something about this, but they shouldn’t.

So here goes nothing. The switch-flipping begins by transferring the endgameviable.com domain to AWS Route 53 control. It will probably take a couple of days to finish completely, based on the tests I did with transferring another domain name.

UPDATE: Something did go horribly wrong. I forgot to set the A record on the domain name, so endgameviable.com was pointing to nothing there for a bit. Then I realized I had the wrong certificate installed, even though I was sure I had installed the right one. Luckily those were the only problems (so far) because, in true professional software development tradition, I didn’t have a plan to go back if it failed. :)

I didn’t setup a lot of monitoring, either, so at this point I have no idea if it’s spewing out thousands of errors per minute or not. I should probably look into that.