k8s

Documentation

Overview

Add code here for finding an outdated pod's controller, issuing rollout-restart to it, and checking progress of the rollout-restart.

We will likely want logic to compare the timestamp of the label we give the pod when marking it as outdated with the status of the controller's rollout-restart.

This way we can figure out if the rollout restart failed for some reason and we should emit a log message and/or metrics indicating the manual intervention might be needed for the pod. We also don't want to issue multiple rollouts for rollouts that are currently progressing.

Index

• func FindPodController(ctx context.Context, kubeClient kubernetes.Clientset, pod corev1.Pod) (*unstructured.Unstructured, error)
• type ControllerNotFoundError
  • func (e ControllerNotFoundError) Error() string
• type PodNotFoundError
  • func (e PodNotFoundError) Error() string

Functions

func FindPodController

func FindPodController(ctx context.Context, kubeClient kubernetes.Clientset, pod corev1.Pod) (*unstructured.Unstructured, error)

I hate this function

Types

type ControllerNotFoundError

type ControllerNotFoundError struct {
	// contains filtered or unexported fields
}

func (ControllerNotFoundError) Error

func (e ControllerNotFoundError) Error() string

type PodNotFoundError

type PodNotFoundError struct {
	// contains filtered or unexported fields
}

func (PodNotFoundError) Error

func (e PodNotFoundError) Error() string