"Severe" Health status after mup reconfig

[markdowney]

First of all, thanks for your work on this package. I've been stuck with the following issue on multiple occasions although it doesn't seem to happen consistently.

After updating my mup config and redeploying, my instances crash to the point where further reconfigs or deploys consistently fail.

I have longEnvVar enabled and 5 instances running. When the instances try to restart, I get the error:
fatal error: An error occurred (404) when calling the HeadObject operation: Key "env/1.txt" does not exist.
Incorrect application version "8" (deployment 2). Expected version "16" (deployment 15).

envVersion defaults to 1 whereas there is no 1.txt in the s3 bucket (largestEnvVersion is 20.txt) and the instances seem to get stuck on this older app version.

What is the correct procedure to troubleshoot this? The problem has happened multiple times. Manually terminating instances via the AWS console has occasionally resolved it but it's been very slow and inconsistent.