AWSAWS Database Migration ServiceDatabase

`aws_dms_replication_instance` cost estimation

A managed instance that runs database migration and replication tasks. Priced per instance-hour by class, doubled for Multi-AZ, plus storage.

An aws_dms_replication_instance is the compute that runs Database Migration Service tasks, moving or continuously replicating data between databases. The migration tasks themselves are free; you pay for the replication instance that executes them.

Cost is per instance-hour by class (dms.t3.micro up to dms.r5.24xlarge), billed continuously while the instance exists. A dms.t3.medium runs about $0.146/hour (~$107/month) in Single-AZ. Enabling Multi-AZ for resilient ongoing replication doubles the instance cost, since it runs a standby. Allocated storage for the instance's logs and cache bills per GB-month on top.

c3x reads replication_instance_class and multi_az and prices instance-hours at 730 hours/month, doubling for Multi-AZ. Storage is priced from allocated_storage. The common waste is leaving a replication instance running after a one-time migration completes, the instance bills until you delete it, so the estimate is a reminder to tear down finished migrations.

Terraform example

A minimal but realistic configuration that C3X can estimate.

resource "aws_dms_replication_instance" "migrate" {
  replication_instance_id    = "prod-migration"
  replication_instance_class = "dms.c5.large"
  allocated_storage          = 100
  multi_az                   = false
  engine_version             = "3.5.2"

  publicly_accessible = false
}

Pricing dimensions

What you actually pay for when you provision aws_dms_replication_instance.

Dimension	Unit	What's being charged
Instance hours	per instance-hour	Per-hour rate for the replication instance class, billed continuously. Multi-AZ doubles it. c3x assumes 730 hours/month. $0.146/hour for dms.t3.medium (Single-AZ)
Allocated storage	per GB-month	Storage for logs and cached changes on the instance. c3x prices allocated_storage. $0.115/GB-month

Sample C3X output

Example output from c3x estimate (dms.c5.large, Single-AZ, 100 GB):

aws_dms_replication_instance.migrate
├─ Instance hours (dms.c5.large)   730  hours        $146.00
└─ Allocated storage               100  GB-month       $11.50

OVERALL TOTAL                                         $157.50

Optimization tips

Common ways to reduce aws_dms_replication_instance cost without changing the workload.

Delete the instance after a one-time migration

Full instance cost post-migration

The replication instance bills continuously until deleted, even after the migration task finishes. For one-time migrations, tear it down as soon as cutover completes.

Use Single-AZ for one-time or non-critical migrations

50%

Multi-AZ doubles the instance cost for standby resilience. A one-time bulk load rarely needs it; reserve Multi-AZ for long-running, business-critical ongoing replication (CDC).

Right-size the instance class to the migration load

Per oversize step

Migration throughput is bounded by the instance class. Start with a modest class and scale up only if the migration is too slow; oversized instances waste money on idle capacity.

FAQ

How does c3x estimate DMS cost?

It reads replication_instance_class and multi_az, prices instance-hours at 730 hours/month (doubled for Multi-AZ), and adds allocated_storage per GB-month. Migration tasks themselves are free.

Are migration tasks charged?

No. DMS tasks (the actual migration/replication jobs) are free. You pay only for the replication instance that runs them and its allocated storage.

Why does Multi-AZ double the cost?

Multi-AZ runs a standby replication instance in another AZ for failover during ongoing replication. Both instances bill, so the cost doubles. It's worth it for critical CDC, not for one-time loads.

What's the biggest source of waste with DMS?

Leaving the replication instance running after a one-time migration finishes. It bills by the hour until deleted. Tear it down at cutover; the task completing doesn't stop the instance charge.