Introduction

Use the attack to drain one or multiple nodes. The nodes are drained via kubectl drain.

Some details:

• Steadybit agents and extensions are excluded from the drain
• kubectl parameter force is used to continue even if there are pods that do not declare a controller
• kubectl parameter ignore-daemonsets is used to ignore daemonsets
• kubectl parameter delete-emptydir-data is used to delete local data of pods using emptyDir

Use Cases

• Check your application failover when a node is drained
• Check if new nodes are created by your autoscaler (in combination with node count check)

Parameters

Rollback

A drained node will be automatically uncorden after the given duration or in case of an error to rollback the effect.