[kn-admin]As a Knative administrator, I want to upate autoscaling configs(global) which apply to overall Knative platform.

[zhanggbj]

A developer may set some autoscaling configs like --limits-cpu or --limits-memory as what kn service create described today. However, a Knative administrator may need to update some overall configs like stable-window or panic-window-percentage or ticker-interval for the whole Knative platform. The kn-admin will list all the config we support to update by kn-admin

According the Knative 0.16 doc about autoscaling, plan support following global configs to help with admin work. (Per-revision config is not planned here, that are more for developer experience)

$ kn admin autoscaling update
- Stable-window xx
- panic-window-percentage xx
- max-scale-up-rate xx
- container-concurrency-target-default xx
- container-concurrency-target-percentage xx
- requests-per-second-target-default xx

If there're any other autoscale configs that should be controlled by Knative administrator, please feel free to leave a comment in this issue. Thanks!

[zhanggbj]

CC @maximilien @julz @duglin ^^^^^^

[zhanggbj]

/assign

[duglin]

I think you're missing:

• target-burst-capacity
• panic-threshold-percentage
• max-scale-down-rate
• enable-scale-to-zero
• scale-to-zero-grace-period
• scale-to-zero-pod-retention-period
• pod-autoscaler-class
• activator-capacity

Basically, everything in here: https://github.com/knative/serving/blob/66c2885ea639baa8fbdbe278e310b7b68d530c39/pkg/autoscaler/config/config.go#L43

[zhanggbj]

@duglin Thanks Doug for pointing this out!
I already added the support to enable and disable scale-to-zero.

• enable-scale-to-zero

So in this PR I'll implement following flags for autoscaling global settings. The default setting will be consistent with values in knative doc

• Stable-window xx
• panic-window-percentage xx
• max-scale-up-rate xx
• container-concurrency-target-default xx
• container-concurrency-target-percentage xx
• requests-per-second-target-default xx
• target-burst-capacity
• panic-threshold-percentage
• max-scale-down-rate
• scale-to-zero-grace-period
• scale-to-zero-pod-retention-period
• pod-autoscaler-class
• activator-capacity