AWSAmazon ElastiCacheDatabase

`aws_elasticache_cluster` cost estimation

Managed Redis or Memcached. Priced per node-hour by instance type, with no storage or request charges.

An aws_elasticache_cluster is a managed in-memory cache running Redis or Memcached. Unlike S3 or DynamoDB, ElastiCache has no per-request or per-GB-month charges. You pay for the underlying compute nodes by the hour, period.

The cost model is identical to EC2: pick a node type (cache.t4g.micro through cache.r7g.16xlarge), pick a region, and pay the hourly rate times the number of nodes. A 3-node cache.m6g.large cluster in us-east-1 costs about 3 × $0.156/hour × 730 hours = $342/month.

Redis clusters have two replication modes. A non-cluster-mode cluster has one primary plus zero or more replicas (set via num_cache_nodes). A cluster-mode cluster shards data across multiple primary nodes, each with optional replicas (aws_elasticache_replication_group, not aws_elasticache_cluster). Pricing is the same per node either way; cluster mode just adds more nodes.

Memcached clusters always use multiple nodes for sharding. Each node is billed independently.

Reserved nodes (aws_elasticache_reserved_cache_node) cut prices by 30-60% for 1- or 3-year commitments. Useful for steady-state caches that you know you'll run continuously.

c3x reads node_type, num_cache_nodes, and engine from the Terraform configuration. Reserved pricing is applied via purchaseOption in c3x-usage.yml.

Terraform example

A minimal but realistic configuration that C3X can estimate.

resource "aws_elasticache_cluster" "session_cache" {
  cluster_id           = "session-cache"
  engine               = "redis"
  node_type            = "cache.m6g.large"
  num_cache_nodes      = 1
  parameter_group_name = "default.redis7"
  engine_version       = "7.1"
  port                 = 6379
}

Pricing dimensions

What you actually pay for when you provision aws_elasticache_cluster.

Dimension	Unit	What's being charged
Node hours	per node per hour	Hourly rate for the node type. Multiplied by num_cache_nodes for the total. c3x assumes 730 hours/month. $0.156/hour for cache.m6g.large in us-east-1
Data transfer between AZs	per GB	Cross-AZ replication traffic for Multi-AZ clusters. Usage-based.
Data transfer out to internet	per GB	Egress to the public internet. ElastiCache is typically accessed from EC2 in the same VPC (free).

Optimization tips

Common ways to reduce aws_elasticache_cluster cost without changing the workload.

Use Graviton (cache.m6g, cache.r6g) over Intel nodes

About 20%

Graviton ElastiCache nodes are roughly 20% cheaper than equivalent Intel nodes (m5, r5) with similar or better performance for Redis and Memcached workloads.

Buy reserved nodes for steady-state caches

30-45%

Caches rarely scale up and down. A 1-year all-upfront RI cuts on-demand cost by 30-45%. Model with purchaseOption: 'reserved' in c3x-usage.yml.

Consider ElastiCache Serverless for spiky workloads

Workload-dependent

Serverless bills by GB-hour of data stored and ECPU operations, not by node-hour. Cheaper for unpredictable or low-baseline workloads. Replaces provisioned clusters with cache.serverless.

Right-size the node type

30-50%

ElastiCache memory utilization is a CloudWatch metric (DatabaseMemoryUsagePercentage). Nodes running below 40% memory can usually drop a size. Check CloudWatch before resizing.

FAQ

Does c3x handle ElastiCache Serverless?

Yes. ElastiCache Serverless is a separate resource type with different pricing (per GB-hour stored, per ECPU). c3x detects the serverless config and uses the right pricing model.

Is Multi-AZ replication free?

The replica node itself costs the same hourly rate as the primary. Cross-AZ replication data transfer is billed per GB but typically small for Redis.

Does c3x estimate Redis storage?

ElastiCache doesn't bill by storage size. The node type defines the memory limit; you pay for the node regardless of how much memory you use. There's no separate storage line item.

What's the difference between aws_elasticache_cluster and aws_elasticache_replication_group?

aws_elasticache_cluster is for single-shard Redis or Memcached. aws_elasticache_replication_group is for Redis with multiple shards (cluster mode) or with explicit primary/replica topology. c3x estimates both, with replication_group being more common for production Redis.