Comment 8 for bug 1856560

cloud-init is used to perform first time configuration of cloud-images using a cloud metadata source. I.e. a generic image is launched by openstack and provisions per-user keys / packages / etc.

If someone manually installs Ubuntu Desktop / Server using an installer, one should not be installing or using cloud-init as clearly, it will never complete without a correct metadata source present. And when it fails to complete, it re-attempts to reprovision the machine on every boot.

If you don't know whether or not there is a cloud metadata source, and whether or not cloud-init should be provisioning the instances on first boot, or whether a desktop/server installers were used - you need to sort that out first. Please escalate to your IT/devops team to understand whether or not it is intentional that cloud-init is installed and attempted on every boot, and where the metadata for it is suppoed to come from. Then fix cloud-init/metadata such that it completes, or remove/disable/uninstall cloud-init.

This is not a bug, but effectively a support issue at this point. The logs are inconsistent with normal cloud-init usage, and there is no reproducer.