Maybe go two weeks' back and you'll get a clean instance, but that's two weeks' data loss, I've seen (non-tech) institutions hit where an hour of data loss is worth paying a ransom for.
https://en.wikipedia.org/wiki/Infrastructure_as_code
https://en.wikipedia.org/wiki/Virtual_machine
https://en.wikipedia.org/wiki/Disk_image
https://en.wikipedia.org/wiki/Shadow_IT
https://en.wikipedia.org/wiki/Von_Neumann_architecture
Separation of code and executables is a nice idea that approximately 0% of organisations fully adhere to.
> "definition files"
Not executable. Text. Readable by humans. Inspectable by humans so you can root out rootkits. Not even the valuable data that cyber criminals go for anyway - they go for personal and financial data, not k8s config files.
> https://en.wikipedia.org/wiki/Virtual_machine
> https://en.wikipedia.org/wiki/Disk_image
Neither of those are relevant. You don't back up virtual machines or image disks - you take afore-mentioned plain-text, audited config files and spin up new instance from scratch.
> https://en.wikipedia.org/wiki/Shadow_IT
If those are actually shadow IT, they won't be in the backups anyway.
> https://en.wikipedia.org/wiki/Von_Neumann_architecture
This is irrelevant snark. If you back up a data file, it doesn't matter that it's stored in the memory of a Von Neumann architecture - it's only going to be used as a data file.
> Separation of code and executables is a nice idea that approximately 0% of organisations fully adhere to.
Citation needed. Also, you just said:
> If the ransomware operators follow best practices
...so are we considering the ideal case, or not?
> I'm really not sure that has a serious answer.
Being snide is bad by itself, but it's even worse when you're wrong on top of it.
> If those are actually shadow IT, they won't be in the backups anyway.
Okay whatever then. I really don't have the energy. I'm just depressed people might believe you.