Open hshabbar opened 3 years ago
@mikestef9 Is the value that EKS uses for this flag documented anywhere?
@mikestef9 Is the value that EKS uses for this flag documented anywhere?
You can check it in Cloudwatch when you enable the control plane logging. In my EKS 1.21 cluster it's: --terminated-pod-gc-threshold="12500"
That's long ago deleted, though I'll take that as its set to the default.
Another flag that should be configurable is --horizontal-pod-autoscaler-tolerance
. The default value is 0.1
, or 10%, but there are use cases to make the value smaller (more sensitive, more responsive) or larger.
+1 we need this feature
+1
Just to add to the list of parameters that we'd like configurable. --concurrent-horizontal-pod-autoscaler-syncs
which is new in K8s 1.26.
Is there like any reason to not allow overriding this --terminated-pod-gc-threshold
default value? it's really a pain to have to clean up those terminated pods manually or having a cronjob to do it...
Any plans of having this feature rolled out anytime soon ? This would be very helpful.
+1 for this very important feature.
+1
Community Note
Tell us about your request What do you want us to build?
Ability to set values for kube controller manager parameters such as --terminated-pod-gc-threshold
Which service(s) is this request for? EKS