Upgrade to a Standard Tier Memorystore for Redis instance with 5 GB capacity and multiple read replicas, then decommission the old instance.

Switch to a Memorystore for Memcached instance with at least three nodes, each with 4 GB memory. Adjust the Dataflow pipeline and clients accordingly, then remove the old Redis instance.

Deploy several new Basic Tier Memorystore for Redis instances, each with 4 GB capacity. Update the Dataflow pipeline and clients to distribute the load across these instances.

Transition to a Standard Tier Memorystore for Redis instance with 4 GB capacity and high availability only (no read replicas), then delete the previous instance.