KEP-4603: Node specific kubelet config for maximum backoff down to 1 …

…second (#128374)

* Add feature gate, API, and conflict validation tests for enablecrashloopbackoffmax

Signed-off-by: Laura Lorenz <[email protected]>

* Handle when current base is longer than node max

Signed-off-by: Laura Lorenz <[email protected]>

* Update pkg/features/kube_features.go

Co-authored-by: Tsubasa Nagasawa <[email protected]>

* Fix indentation

Signed-off-by: Laura Lorenz <[email protected]>

* Follow convention for success test

Signed-off-by: Laura Lorenz <[email protected]>

* Normalize casing, and change field to Duration

Signed-off-by: Laura Lorenz <[email protected]>

* Fix json name and some other casing errors

Signed-off-by: Laura Lorenz <[email protected]>

* Another one I missed before

Signed-off-by: Laura Lorenz <[email protected]>

* Don't clobber global max function

Signed-off-by: Laura Lorenz <[email protected]>

* Change to flat value in defaults.go

Signed-off-by: Laura Lorenz <[email protected]>

* Streamline validation and defaults

Signed-off-by: Laura Lorenz <[email protected]>

* Fix typecheck

Signed-off-by: Laura Lorenz <[email protected]>

* Lint

Signed-off-by: Laura Lorenz <[email protected]>

* Tighten up validation for subsecond values

Signed-off-by: Laura Lorenz <[email protected]>

* Rename field from MaxBackOffPeriod to MaxContainerRestartPeriod

Signed-off-by: Laura Lorenz <[email protected]>

* A few missed references to renames

Signed-off-by: Laura Lorenz <[email protected]>

* Only compare flags in flags test

Signed-off-by: Laura Lorenz <[email protected]>

* Don't mess with SetDefault signature

Nobody messes with SetDefault signature

Signed-off-by: Laura Lorenz <[email protected]>

* Fix stale signature change, and update test data

Signed-off-by: Laura Lorenz <[email protected]>

* Inspect current feature gates at defaulting time

Signed-off-by: Laura Lorenz <[email protected]>

* Don't use the global feature gate for temp usage

Signed-off-by: Laura Lorenz <[email protected]>

* Expose default error, and some comments

Signed-off-by: Laura Lorenz <[email protected]>

* Hint fuzzer for less arbitrary values to FeatureGates

Signed-off-by: Laura Lorenz <[email protected]>

---------

Signed-off-by: Laura Lorenz <[email protected]>
Co-authored-by: Tsubasa Nagasawa <[email protected]>

Kubernetes-commit: 7fe41da5221b215d097e930c6b7aa52e88324f66

config/v1beta1/types.go

@@ -775,6 +775,11 @@ type KubeletConfiguration struct {
 	// +featureGate=GracefulNodeShutdownBasedOnPodPriority
 	// +optional
 	ShutdownGracePeriodByPodPriority []ShutdownGracePeriodByPodPriority `json:"shutdownGracePeriodByPodPriority,omitempty"`
+	// CrashLoopBackOff contains config to modify node-level parameters for
+	// container restart behavior
+	// +featureGate=KubeletCrashLoopBackOffMax
+	// +optional
+	CrashLoopBackOff CrashLoopBackOffConfig `json:"crashLoopBackOff,omitempty"`
 	// reservedMemory specifies a comma-separated list of memory reservations for NUMA nodes.
 	// The parameter makes sense only in the context of the memory manager feature.
 	// The memory manager will not allocate reserved memory for container workloads.
@@ -975,6 +980,15 @@ type MemorySwapConfiguration struct {
 	SwapBehavior string `json:"swapBehavior,omitempty"`
 }
 
+type CrashLoopBackOffConfig struct {
+	// maxContainerRestartPeriod is the maximum duration the backoff delay can accrue
+	// to for container restarts, minimum 1 second, maximum 300 seconds. If not set,
+	// defaults to the internal crashloopbackoff maximum (300s).
+	// +featureGate=KubeletCrashLoopBackOffMax
+	// +optional
+	MaxContainerRestartPeriod *metav1.Duration `json:"maxContainerRestartPeriod,omitempty"`
+}
+
 // +k8s:deepcopy-gen:interfaces=k8s.io/apimachinery/pkg/runtime.Object
 
 // CredentialProviderConfig is the configuration containing information about