DevOps Glossary

Kubernetes Taints and Tolerations are scheduling controls that help you keep pods away from specific nodes unless those pods explicitly allow the match. A taint is applied to a node as a key, value, and effect, such as NoSchedule, PreferNoSchedule, or NoExecute. A toleration is added to a pod spec so the Kubernetes scheduler may place that pod on a matching tainted node. Teams use them to reserve nodes for system workloads, isolate GPU or high-memory workloads, protect dedicated tenant nodes, or move pods away from unhealthy nodes. They are useful in self-managed clusters and managed services such as Azure Kubernetes Service, but they do not force a pod onto a node by themselves; for that, you usually combine them with node selectors, node affinity, or workload placement rules as part of broader Kubernetes infrastructure management.