Just a simple hole renders them useless. The only method to reconstruct them from there would be any kind of SEM or AFM which would still take weeks to months to years depending on the size/density of the drives.

Even just opening them up and smacking the disks would be sufficient

Next time just encrypt them.

Just because there is no update does not mean there are security vulnerabilities to worry about, or do you have a specific one that is not fixed?

The attack vector seems very narrow to me. It checks the container registry downloads the containers and runs some docker commands.

It has no interface, so in order to attack it you either have to compromise the container registry (but then it would be easier to compromise the containers you download) the secure connection used to download the containers (https is quite stable) or something on the server side.

Also the project does not really look that abundant to me.

EDIT: So i have not checked this, but watchtower is probably using docker for most steps anyway? So basically the only thing that could be attacked is via the notifications watchtower is sending?

Years out of date

What problems does it have? Never ran into an issue for my usecase.

Automatic updates. Works like a dream. Depending on what you are running it can obviously cause issues, either server side breaking or server,client communication issues

You have basically two options.

1. Symmetric Encryption. That means you use the same password/key for writing the Backup and for reading the backup. Here you have to write the password somewhere, depending on the OS there are options like keychains or similar that can hold the password so that the password is only available once you are loged in or have unlocked the keychain.

2. Asymmetric Encryption. That means you have different passwords/keys to read and write the backup. PGP is an example here. Here you can just simply use one key to write the backup, this key can become public and you do not have to worry about your backup since it will only be readable with the 2. key.

I personally use Restic with a password that is only readable by the system root user stored on the filesystem. Since I use Full Disk Encryption i do not have to worry too much about when the secret is available in clear text at runtime.