amazon-memorydb-pricing-calculatoraws-memorydb-pricing-calculator
amazon-memorydb-pricing-calculatoraws-memorydb-pricing-calculatormemorydb-pricing-calculator
Amazon MemoryDB Pricing Calculator - Valkey & Redis OSS
Amazon MemoryDB pricing calculator for Valkey and Redis OSS. Estimate node hours, data written, snapshot storage, and Multi-Region replication costs by AWS Region.
Frequently Asked Questions
How is Amazon MemoryDB node pricing calculated?
MemoryDB bills every node in the cluster per node-hour, at a rate that varies by Region, engine, and node type. Total nodes per Region equals shards multiplied by one plus the replicas per shard, so a 3-shard cluster with 2 replicas per shard runs 9 billable nodes. Replicas are billed at the same rate as primaries.
Is MemoryDB for Valkey cheaper than Redis OSS?
Yes. Across every Region in the current AWS Price List, Valkey node hours are published at 70% of the Redis OSS rate for the same node type. Valkey also includes a monthly free data-written allowance, while Redis OSS data written is charged per GB from the first GB. The two engines are separate published dimensions, so their rates are never interchangeable.
What is the MemoryDB data written charge?
MemoryDB is a durable database that writes to a multi-AZ transaction log, and it charges per GB of data written to the cluster. Reads are not charged. Valkey publishes a monthly free allowance before its paid per-GB tier; Redis OSS is charged at a flat per-GB rate with no regional free tier.
How do MemoryDB reserved nodes work?
Reserved nodes commit to a one-year or three-year term with a No Upfront, Partial Upfront, or All Upfront payment option. AWS publishes an upfront fee and a discounted hourly rate for each combination. Reserved capacity is billed for every hour of the term regardless of utilization, and each Region is reserved independently. This calculator estimates On-Demand cost only, so treat its total as the un-discounted baseline you would reserve against.
Can I buy reserved nodes for MemoryDB for Valkey?
AWS publishes MemoryDB reserved node prices under Redis OSS SKUs only. A Redis OSS reservation can still cover running Valkey nodes through normalization units, and because Valkey carries a lower normalization factor, one reservation covers more Valkey capacity. Because there is no published Valkey reserved rate to price against, this calculator does not model reserved terms at all rather than synthesizing one.
How is MemoryDB snapshot storage billed?
AWS includes a free snapshot allocation with every cluster and charges per GB-month only for storage beyond it. Enter the primary Region’s additional billable amount rather than total snapshot size. Multi-Region estimates exclude secondary snapshots; estimate those separately with the secondary selected as primary.
What does MemoryDB Multi-Region cost?
MemoryDB Multi-Region is a Valkey feature. It adds a full set of nodes in the secondary Region, priced at that Region’s own published node rate, plus inter-Region data transfer out for replicated writes. This calculator prices each Region independently and takes replicated transfer out as an explicit input, because AWS does not publish a fixed ratio between data written and replicated bytes.
Should I add shards or replicas to a MemoryDB cluster?
Each shard owns a slice of the keyspace and has exactly one primary node; replicas add read capacity and failover targets within a shard. Adding a shard raises capacity and cost linearly, while adding a replica multiplies the node count of every shard at once. Going from one to two replicas per shard raises node spend by 50%, so replica count is the single biggest lever on a MemoryDB bill after node size. One replica per shard is usually the cost-effective floor for a production cluster.
When is a MemoryDB data tiering node type cheaper?
The r6gd families pair memory with local NVMe SSD so MemoryDB can tier colder keys to disk. They carry a higher node-hour rate than the equivalent r6g node but hold considerably more data per node, which usually beats scaling out on memory-only nodes for large datasets with a warm working set. The node picker shows published SSD capacity alongside vCPU and memory.
How do I reduce Amazon MemoryDB costs?
Start with engine choice: Valkey is published at a lower node rate than Redis OSS and adds a free data-written allowance, so it is the cheaper starting point at identical capacity. Next, size replicas deliberately, since a replica costs a full node. Then check whether a data-tiering family holds your dataset on fewer nodes. Finally, commit only what you can predict, because reserved nodes bill for the entire term whether or not the cluster runs.
Which charges are excluded from this MemoryDB estimate?
Reads, inbound data transfer, and data transfer between MemoryDB and other AWS services in the same Region are not MemoryDB charges. General internet data transfer out and cross-AZ transfer are also out of scope, since they are not MemoryDB Multi-Region pricing dimensions. Reserved node terms and the restricted account-level Redis OSS data-written allowance are excluded too: the allowance is a limited cross-Region SKU rather than a per-Region tier, so applying it to a single cluster would understate the bill.
Related tools
CloudBurn
Catch AWS cost mistakes before they ship.
Use the calculator for quick estimates, then use CloudBurn when you need a deterministic cost review workflow. Run
scan against Terraform and CloudFormation before deploy, then run discover against live AWS to find the waste that is already burning.Read the docs to learn how it works.