Before you start: migrate from Google Memorystore to KiviDB

Check what the instance uses

gcloud redis instances describe my-redis --region=us-central1
gcloud redis clusters describe my-cluster --region=us-central1
gcloud memorystore instances describe my-valkey --location=us-central1

Look at the version (redisVersion, engineVersion), the tier, the shard count or mode (cluster mode), whether AUTH or IAM authentication is on, whether in-transit encryption is on, and the configs (redisConfigs, engineConfigs) such as maxmemory-policy.

Check	Why it matters
Valkey 9.0 or 9.1 (9.1 is the default for new Valkey instances)	Neither the backup file nor DUMP loads yet: copy with RIOT (below)
Cluster or cluster mode enabled	One RDB per shard; and the client the application uses afterwards
EVAL, EVALSHA, FUNCTION in your code (BullMQ, Sidekiq, many lock libraries)	KiviDB needs the lua or full build; on Cloud, turn on Lua scripting when you create the database
In-transit encryption	The tls or full build; on Cloud, TLS on
IAM authentication	The application needs a password for KiviDB
maxmemory-policy	Set the same one in KiviDB; KiviDB's default is noeviction
Databases other than 0	KiviDB has 16 databases, as Redis does; a sharded KiviDB has only database 0

Size it

In Cloud Monitoring, read the memory in use: redis.googleapis.com/stats/memory/usage (Memorystore for Redis), redis.googleapis.com/cluster/memory/total_used_memory (Redis Cluster), or memorystore.googleapis.com/instance/memory/total_used_memory (Valkey). Take the peak over the last month and add headroom. KiviDB Cloud seeds a database from a file only when the file is no larger than the database.