Our serverless WordPress project is now ended, and you’ve been redirected to our “staging” website for the project, running on an Oakley Studio managed EC2 Instance at Amazon Web Services.
Our serverless website was up and running for about a month. During that time we observed there were brief periods when the site was showing “503 Temporary Interruption in Service” errors, typically following an edit or a plugin update. Our best guess is these interruptions happened when the “serverless” website was syncing either the file system (EFS) or the database (RDS) between availability zones. Or possibly it was an effect of the auto-scaling that happens as traffic increases, or decreases.
In the end, we took down the site after a failed plugin update caused a “504 Bad Gateway” error, and we did not want to spend much time troubleshooting.
We have continued to learn more about the “serverless” and “containerized” approach to website hosting, especially for WordPress. While our “Phase 1” of this project used a Docker container implementation, we anticipate a “Phase 2” in the near future, when we will be trying out Amazon’s Elastic Kupernetes Service (EKS) for our WordPress container.
Check back in another month to find out what we’ve learned!