Paul Nicholson

As a workaround, we set [`max_retries`](https://developer.hashicorp.com/terraform/language/settings/backends/s3#max_retries) on the s3 backend to a value of `1`, which appears to resolve the issue for us at the moment. This was based on...

Tested this in a large cluster where we had frequent oomkill events and can confirm it has helped a lot! Any reason we wouldn't want to default this option to...

Would also find this useful, particularly if it could filter the changelog of commits based on the path (e.g `app-a/`) to determine the next version (rather than all commits between..)