# Inactive MemoryDB Cluster | Cloud Efficiency Hub Canonical: https://www.pointfive.co/efficiency-hub/inefficiencies/inactive-memorydb-cluster MemoryDB clusters are billed for every node, primary and replica, for every hour the cluster exists, whether or not any client is connected. By: PointFive Updated: 2026-09-28 [Cloud Efficiency Hub](https://www.pointfive.co/efficiency-hub) The short version MemoryDB clusters are billed for every node, primary and replica, for every hour the cluster exists, whether or not any client is connected. PointFive Research Cloud cost research at PointFive AWS service [AWS MemoryDB](https://www.pointfive.co/efficiency-hub/cloud-services/aws-memorydb) Category [Databases](https://www.pointfive.co/efficiency-hub/service-category/databases) Reference CER-0374 Type Idle or Unused Resource ## Explanation Why the waste happens and who it affects. Clusters created for a proof of concept, a load test, a migration rehearsal or a feature that was later retired often keep running after the application stops using them, and because MemoryDB is typically deployed with replicas across Availability Zones for durability, an unused cluster usually means several idle nodes rather than one. The waste is easy to miss because the cluster stays healthy and quiet: no errors, no alarms and no traffic, only a steady node-hour charge. It is most common in non-production accounts and in teams that evaluated MemoryDB alongside ElastiCache or another data store and kept both. AWS Compute Optimizer classifies MemoryDB clusters as idle using connection, CPU and keyspace activity. ## Billing model The pricing dimensions that drive this cost. MemoryDB charges for provisioned nodes regardless of whether the cluster receives any traffic. On-Demand node-hours Billed per node per hour based on node type, for every primary and replica node in the cluster while it exists Data written Billed per GB written to the cluster, with a monthly free allowance for the Valkey engine Snapshot storage Snapshot storage up to 100% of the cluster's total storage is included, and storage beyond that is billed per GB-month Reserved nodes 1- or 3-year commitments billed for every hour of the term whether or not a matching node is running ## How to detect 5 checks to find it in your estate. - Review AWS Compute Optimizer idle recommendations for Amazon MemoryDB, which flag clusters in available status that had no new connections, engine CPU utilization below 1% and no keyspace hits or misses over the 14-day lookback period - In CloudWatch (AWS/MemoryDB namespace), check NewConnections, CurrConnections, KeyspaceHits, KeyspaceMisses, GetTypeCmds and SetTypeCmds per node over at least several weeks; remember CurrConnections includes a few connections MemoryDB uses to monitor the cluster, so a small constant value does not indicate application use - Check whether any application configuration, secret or infrastructure-as-code module still references the cluster endpoint - Identify the owner through tags and creation context (proof of concept, migration, test environment) and confirm whether the cluster is still required - Check Cost Optimization Hub, which includes MemoryDB clusters among its supported resources, for an estimated savings figure after discounts ## How to fix 5 ways to remove the waste. - Delete clusters confirmed as unused; pass --final-snapshot-name to delete-cluster (or choose Yes for a final snapshot in the console) if the data may be needed later, since automatic snapshots are not retained after deletion - Delete old manual snapshots once they are no longer needed, because manual snapshots are retained after the cluster is deleted and can continue to incur storage charges beyond the included allowance - For non-production clusters that must remain, reduce the number of replicas per shard or move to a smaller node type to cut the node-hour charge - If reserved nodes were purchased for the cluster, check whether another MemoryDB cluster in the same Region can use the reservation before deleting, as the reservation continues to be billed for the full term - Add ownership tags and a review date to new MemoryDB clusters so temporary clusters are cleaned up on schedule ## Documentation Vendor references for pricing and configuration. - [Pricing for Amazon MemoryDB aws.amazon.com](https://aws.amazon.com/memorydb/pricing/) - [Viewing idle resource recommendations docs.aws.amazon.com](https://docs.aws.amazon.com/compute-optimizer/latest/ug/view-idle-recommendations.html) - [Metrics for MemoryDB docs.aws.amazon.com](https://docs.aws.amazon.com/memorydb/latest/devguide/metrics.memorydb.html) - [Creating a final snapshot docs.aws.amazon.com](https://docs.aws.amazon.com/memorydb/latest/devguide/snapshots-final.html) ## Related inefficiencies [Browse the library](https://www.pointfive.co/efficiency-hub) - AWS MemoryDB CER-0166 ### [Suboptimal Engine Selection in MemoryDB](https://www.pointfive.co/efficiency-hub/inefficiencies/suboptimal-engine-selection-in-memorydb) MemoryDB now supports Valkey, a drop-in replacement for Redis OSS offering significant cost and performance advantages. However, many deployments still default to Redis OSS, incurring higher hourly costs and unnecessary data write charges.... Databases - AWS RDS CER-0299 ### [Orphaned RDS Backup Storage After Database Deletion](https://www.pointfive.co/efficiency-hub/inefficiencies/orphaned-rds-backup-storage-after-database-deletion) An RDS database instance is deleted but its manual snapshots or retained backups remain. Unlike automated backups tied to a live instance, these backups persist independently and continue generating storage costs despite no longer... Databases - AWS Aurora CER-0293 ### [Automatic Restart of Stopped Aurora Clusters Causing Unintended Compute Charges](https://www.pointfive.co/efficiency-hub/inefficiencies/automatic-restart-of-stopped-aurora-clusters-causing-unintended-compute-charges) Amazon Aurora database clusters are intentionally stopped to avoid compute costs but are automatically restarted by the service after the maximum allowed stop period. Once restarted, re-started database instances begin accruing... Databases --- Source: the public page above. Product screenshots and illustrative interfaces are examples, not live customer data.