Ask HN: Backup and Restore Plans

We've been debating at work what level of backing up is safe, and at what point is it overkill.

We run our own hardware and use VMWare to logically slice and dice that up. Our software runs in containers on vanilla CentOS VMs. The containers are orchestrated with k8s. Our databases use full replication with hot stand-bys and we take full dumps twice daily. Really nothing fancy or complicated.

I am under the impression that maintaining our normal code, docker configs, k8s configs, ansible playbooks (to setup the docker and k8s daemons and insert ssh keys, etc) in git repos, and routine database snap shots is enough. However I have a coworker advocating for full VM snap shots as well to aid in recoveries.

Are VM snap shots overkill, or am I being too reliant on, and idealistic about, our ability to maintain good automation?