The DevOps way is to have them die at regular intervals in addition to other triggers and then rebuild on a regular cadence. Iirc correctly Netflix servers have a 12 hour TTL. Windscribe could easily do a 1-2 hour TTL with matching certs and encryption keys.
oftenawake@lemmy.dbzer0.com 20 hours ago
Surely the servers aren’t running on bare metal anyway? So wouldn’t they just keep the virtual servers in ram and destroy them regularly to redeploy from an image? (I have no idea - I was a web dev 20 years ago!)
But it seems like when you have imaginary “computers” that can be regularly destroyed and rebuilt at little cost or hassle, there shouldn’t be much point in trying to capture or examine the actual hardware because all it’s doing is managing virtual machines existing in ram?