jcscottiii
commented
2 years ago @KyleJu that's correct. After terraform apply happens, it records the state of things in a state file. Right now, the state file is stored in the repository. But terraform can be configured to store the state file in remote blob storage (e.g. google cloud storage). That would prevent these commits. But we need to turn on versioning and pay for that. It wouldn't be a big cost. But definitely a discussion point. If we plan on updating this a lot, I would highly suggest moving there. But since we don't do it often, I am 100% fine leaving it as-is.
Following the instructions for deploying in the README. terraform apply deployed the new docker images.
This was kept separate from the OS updates in #60. That updated the OS version image. (Wanted to minimize the amount of changes at once) This PR updated the docker images for cert renewers and the wpt.live server
All of these changes were generated automatically by terraform upon apply
The reason why there was no terraform configuration change and only terraform state changes: The terraform just pulls the latest docker image in the registry. (After doing the instructions in the README, it pushes new versions to the docker registry) We might want to pin that to certain versions in case we need to do a rollback.