Some critical workloads require continuous high availability and should not be blocked by node taints in the Kubernetes cluster. By applying a universal toleration to your pods, you can ensure that these pods are scheduled regardless of the taints present on the nodes.

To allow pods to tolerate any or all taints, configure them with a toleration using the operator: "Exists" parameter and no key or value. This configuration means the pod will not be restricted by any taints applied to nodes.

Add the following section to your pod specification:

spec:
  tolerations:
  - operator: "Exists"

Outcome

Critical pods configured with this universal toleration can run on any node without being affected by taints, ensuring maximum availability.