Apologies in advance if this has already been discussed.
I’ve searched through the forums and watched this tutorial but I’m still looking for some additional help. I think this may warrant a feature request.
Coming from Pterodactyl, I relied heavily on the ability to bind/pin instances to specific CPU cores. This was critical for isolating workloads.
The forum post I linked is two years old, and @Mike at the time, your response might have unintentionally conflated the two technologies? Please forgive my directness here, but for clarity: CPU shares adjust the relative weight where-as CPU pinning binds containers to specific cores. Pinning is intrinsically important to me as it eliminates CPU scheduling competition from other workloads, and provides deterministic performance. If AMP doesn’t support CPU pinning directly, I’d like to propose it as a feature request, please. 
My other topic of discussion is on backups.
How does AMP itself safely backup? Normally you would want to stop an application from running.
Currently I don’t have AMP scheduled to backup. My first run with it, I experienced an OOM event so I’ve been leery of trying it again.
Where are the backups going by default, can this be changed to another local path?
If I do docker, I’m concerned how that would effect backing up and accessing the files.
Does AMP docker use traditional volume paths? /var/lib/docker/volumes
Could I just use a bind mount maybe?
Currently, I’m using back in time to backup my AMP home directory, which I think is unsafe for data integrity. I think if I better understood AMP’s own backup feature I might go back to that.. Perhaps I’m overlooking documentation that sheds light on backups.
I’m very grateful for any assistance on these matters.