-
Notifications
You must be signed in to change notification settings - Fork 13
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Resource requests and limits #22
Comments
Here's the possible formats for Pod CPU resource requests and limits:
All formats can be present in the same Pod spec and are directly translatable to milliseconds, which is supported by
Do you see any limitation of just going w/ that attribute? |
[ Quoting <[email protected]> in "Re: [virtual-kubelet/systemk] Resou..." ]
Here's the possible formats for Pod CPU resource requests and limits:
• an integer value, such as 2
• a fraction, such as 0.1
• an integer value in millis, such as 500m
All formats can be present in the same Pod spec and are directly translatable
to milliseconds, which is supported by CPUQuotaPeriodSec, eg:
• 2 == 2.0 == 2000m
• 0.1 == 100m
Do you see any limitation of just going w/ that attribute?
nope, if you say these are compatible it's pretty easy to convert and use.
|
Actually, I got it wrong when it comes to systemd primitives. Now, this only covers resource limits and not requests. |
[ Quoting <[email protected]> in "Re: [virtual-kubelet/systemk] Resou..." ]
Actually, I got it wrong when it comes to systemd primitives. CPUQuota is the
field to use and it supports percentage alone. The conversion should be
trivial, though.
Now, this only covers resource limits and not requests.
don't think we can do requests though?
We'll also need to enable the CPU and memory accounting
|
One problem I see is that it seems that enabling accounting for one slice means parent and child slices are affected too which may not be desirable. |
systemd has both mem and cpu accounting. The memory is specified in bytes and can be used from the podspec as is. The cpu however is done differently than in k8s. See https://www.freedesktop.org/software/systemd/man/systemd.resource-control.html
The problem is how to translate the cpu in the podspec into systemd values.
The text was updated successfully, but these errors were encountered: