This page explains how to scale instances, and how scaling works for Memorystore for Memcached.
You can scale the number of nodes in a cluster down to a minimum of 1 node or up to a maximum of 20 nodes. You cannot scale a node's memory capacity or number of vCPUs.
When you scale up your cluster, unless you use Auto Discovery you must manually add the IP addresses of the new nodes to your application so the application can access them. Also, Memorystore for Memcached does not rebalance the data in your nodes. You must manage any data rebalancing you wish to perform.
If you scale down your instance, you will lose the data that is stored in the nodes that are deleted from your cluster.
During a scale down operation, Memorystore removes nodes from the end of the node list. For example, if you are scaling from 12 nodes to 8 nodes, nodes 9-12 will be deleted from the instance.
You can view the node list on your Memcached instance's node list tab in the Cloud Console.
You can also view the node list by running the following command, replacing variables with appropriate values:
gcloud beta memcache instances describe instance-id --region=us-region
Managing nodes when scaling down
Your application's node management strategy affects what actions you need to take when scaling down your cluster.
If you manually manage the addition and removal of Memcached nodes in your application, you must adjust your code by removing references to the nodes that will be deleted before scaling down the cluster.
If you use Auto Discovery, node addition and removal is handled for you automatically.
Scaling the number of nodes in your cluster
Go to the Memorystore for Memcached page in the Google Cloud Console.
Click the Instance ID of the instance you want to resize.
In the Nodes field, enter your desired number of nodes.
Enter the following command, replacing the variables with the appropriate values:
gcloud beta memcache instances update instance-id --region=region --node-count=desired-number-of-nodes
gcloud beta memcache instances update myinstance --region=us-central1 --node-count=6