Manually scaling deployments

With autoscaling available, there are limited reasons why you would want to scale the size of your container clusters and GKE node pools manually. Some scenarios that come to mind are testings for optimal resource allocation and production scenarios requiring high levels of resources immediately, where scale-up time would be a detriment.

You can update your GKE cluster size via the cloud console by editing the appropriate node pool. You can complete the same resize operation using the gcloud CLI:

gcloud container clusters resize CLUSTER_NAME --node-pool NODE_POOL --size SIZE

To manually scale your pods, use the Kubernetes CLI to set the static replica size. This is a single static size, whereas when auto scaling, we will define a minimum and maximum number of replicas:

kubectl scale [CONTROLLER] NAME --replicas SIZE