Kubernetes 1.25: Stability for StatefulSets and DaemonSets in Rollout

Introduction

With the release of Kubernetes 1.25, there are exciting updates for developers and DevOps teams: Two important features, minReadySeconds for StatefulSets and maxSurge for DaemonSets, have reached “stable” status. These changes improve the handling of rollouts and increase application availability.

What specifically changes for developers/DevOps teams?

The new stable features offer developers enhanced control over the rollout process of StatefulSets and DaemonSets. minReadySeconds allows teams to ensure that a Pod is ready for a specified time before it is considered available. At the same time, maxSurge enables multiple instances of a Pod to run on a node during a rollout, minimizing downtime.

Practical examples or use cases

minReadySeconds for StatefulSets

Consider running a StatefulSet for a database-backed application. With minReadySeconds, you can ensure that each Pod has enough time to initialize—such as loading cache data—before it is reported as available. This is particularly useful if your application relies on quick response times. Here’s an example of how to configure minReadySeconds:

apiVersion: apps/v1
kind: StatefulSet
metadata:
  name: my-application
spec:
  updateStrategy:
    type: RollingUpdate
  minReadySeconds: 10
  ...

maxSurge for DaemonSets

Imagine using DaemonSets for system-level components like CNI or CSI plugins. With maxSurge, you can ensure that additional Pods run during a rollout to maintain availability. This means that even during updates, no one has to forego essential network services. Here’s how to set up maxSurge:

apiVersion: apps/v1
kind: DaemonSet
metadata:
  name: my-daemonset
spec:
  updateStrategy:
    type: RollingUpdate
    rollingUpdate:
      maxSurge: 2
  ...

These new features in Kubernetes 1.25 are not just technical improvements but also contribute to increasing the stability and reliability of your applications. At ayedo, we are excited to support you as a Kubernetes partner and help you implement these exciting new features.

Source: Kubernetes Blog