[VC-32659] Remove the default CPU limit#539
Conversation
Signed-off-by: Richard Wall <richard.wall@venafi.com>
Signed-off-by: Richard Wall <richard.wall@venafi.com>
hawksight
left a comment
There was a problem hiding this comment.
Having read through I think I agree.
Ultimately the choice is down to the platform team and their policy around what to set and not set.
The bonus here is that we are less likely to have a new install in a large cluster killed for resource starvation (of CPU) and require debugging.
|
@tfadeyi Hey, I'm not sure whether this needs to be reviewed by you, WDYT? |
maelvls
left a comment
There was a problem hiding this comment.
Thanks for making this change, it's a good idea to not set CPU limits by default... especially because none of the "pros" for using CPU limits apply to the Venafi Agent: avoiding reliance on unpredictable spare cpu cycles due to low cpu requests; getting more predictable QoS; being able to charge back on multi-tenant environment.
|
LGTM I don't think it requires my approval, I think @wallrj should have permission to merge |
In https://docs.venafi.cloud/vaas/k8s-components/c-k8s-components-best-practice/#cpu we say:
So I've removed the default CPU limit from the chart values for venafi-kubernetes-agent.
I have not changed the
jetstack-agentcharts, because I think it is now considered legacy.I also removed the note about using vertical pod autoscaler, because it seems irrelevant. We don't mention VPA in the documentation or in the other charts AFAIK.