Open
Description
Is there an existing issue for this?
- There is no existing issue for this feature
What are you currently unable to do
Have a say in what the OOM killer kills
What do you think would need to be added
I had some VMs randomly in a stopped state and it turned out that my host went OOM :(
Unfortunately when this happens random VMs and LXCs are getting killed.
I would like to request the ability to set an OOM score per instance, as some of my instances are more important than others.
incus config set <instance> limits.oom_priority <value>
Something like this ranging from -1000 to 1000, where a low value is never killed and a higher value is more likely to be killed. Similar to OOMScoreAdjust
from systemd.