Thought it is not recommanded, you can upgrade the machine type used by your Cloud Composer cluster by removing and create again de default-pool
.
There's an official documentation on how to to this.
It might not show in the Composer UI or in the Composer settings returned by CLI.
Also, you can as suggested in parakeet's comment use a KubernetesPodOperator
or GKEPodOperator
in order to properly dimension the node pool used to instanciate the pod.
与恶龙缠斗过久,自身亦成为恶龙;凝视深渊过久,深渊将回以凝视…