The Startup Script doesn't run as it exists today in the gcve-monitoring module. When this doesn't run, the health check fails. When that fails, the instance is replaced over and over. When I logged in and ran sudo journalctl -u google-startup-scripts.service, I get the message shown in the screenshot here.

The key is "/usr/bin/env: bash\r : No such file or directory and then the script ends with exist status 127. The odd thing is that is I manually copy and run the startup script, it runs as expected.

Upon further investigation, it appears to be an issue when run from a Windows Server, but works OK from CloudShell (and presumably any other Linux environment). Maybe update the Readme with the requirement to run from Linux or update the editor (such as VS Code) to handle correctly (just in Linux) or run the git config core.autocrlf false before cloning the repository.