Node hours for a managed Redis or Memcached cache cluster.
Managed Redis bills every hour the node exists, so an oversized cluster or an unused replica is pure waste. Size it to the working set and reserve for steady load. Pair with the managed database calculator.
This ElastiCache / Redis Cost Calculator estimates your monthly cache bill from three inputs: the total number of nodes (your primary plus any replicas), the price per node per hour for your chosen node type, and the hours per month the cluster runs. It multiplies these together β nodes Γ hourly rate Γ hours β to project what a Redis or ElastiCache deployment costs before you provision it. The main drivers are node count and node-type pricing: adding replicas for read scaling or high availability multiplies the hourly rate across every node, so a small change in either input moves the total noticeably.
The key trade-off is replicas versus spend. Each replica improves availability and read throughput but adds a full node's hourly cost, and because cache clusters typically run 24/7 (about 730 hours a month), that cost is continuous rather than occasional. Before committing, compare a smaller number of larger nodes against more small nodes at the same total memory, and confirm you actually need every replica β right-sizing the node type is usually the cheapest lever.
Per node-hour, based on the instance type and its memory. A primary plus one replica for high availability means you pay for two nodes around the clock. There is no per-request charge β you pay for the running memory.
Right-size the node memory to your working set, drop replicas you do not need for availability, and buy reserved nodes for long-lived clusters. A cache that is far bigger than the hot data is just paying for empty RAM.