Estimating costs

Before you commit to a large upload, cargoship estimate tells you what it will cost to store on S3 — across storage classes, with CargoShip's compression and chunking factored in. It's read-only and touches no data, so run it as often as you like.

cargoship estimate ./my-data

CargoShip scans the directory, models how it will compress and chunk, and prints monthly storage, request, and transfer costs. For the full flag list, see the cost command reference.

Compare naive upload vs. CargoShip

The most useful thing estimate does is show what CargoShip's chunking and compression save you versus uploading files as-is:

cargoship estimate ./my-data --show-comparison

This contrasts a "naive" upload (one S3 object per file, uncompressed) against CargoShip's approach (compressed tar.zst chunks). For datasets with many small files, the difference in request costs alone is often dramatic — thousands of PUT requests collapse into a handful of chunk uploads.

Why request count matters

S3 bills per 1,000 PUT/GET requests. Ten thousand tiny files uploaded individually is ten thousand PUTs; the same files packed into a dozen chunks is a dozen PUTs. --show-comparison makes that visible before you pay for it.

Pick a storage class

cargoship estimate ./my-data --storage-class GLACIER

Pass any S3 storage class to see its monthly cost. By default the estimate also prints recommendations across classes so you can weigh storage price against retrieval cost and latency. Common choices:

Class	Best for	Relative storage cost
STANDARD	Frequent access	Baseline
INTELLIGENT_TIERING	Unknown / mixed access	Auto-tiers
STANDARD_IA	Monthly access	Lower
GLACIER_IR	Archive, occasional instant retrieval	Much lower
GLACIER	Cold archive (minutes to retrieve)	Very low
DEEP_ARCHIVE	Compliance, rarely retrieved (hours)	Lowest

See Lifecycle & storage classes for choosing between them and automating transitions.

Real-time pricing

By default, estimates use CargoShip's built-in pricing table (fast, no credentials needed). To pull current rates from the AWS Pricing API for your exact region:

cargoship estimate ./my-data --storage-class GLACIER_IR --real-time-pricing --region us-west-2

Requires AWS credentials

--real-time-pricing calls the AWS Pricing API, so valid credentials and network access are required. Without the flag, CargoShip falls back to its bundled pricing — accurate enough for planning, but not penny-exact for negotiated or enterprise rates.

JSON output for scripts

cargoship estimate ./my-data --storage-class GLACIER_IR --format json

Machine-readable output for wiring estimates into CI gates, dashboards, or approval workflows. The default --format table is meant for humans.

Other options

• --region — region for pricing (default us-east-1); match your target bucket.
• --bandwidth — network bandwidth in MB/s to model upload time (0 = auto-detect).
• --show-recommendations / --show-parallel / --show-upload-optimization — on by default; each adds an advisory section (storage-class picks, parallel-upload tuning, chunk sizing). Disable any you don't want in scripted output.

Best practices

TIP

• Always estimate before a first large upload — it's free and takes seconds.
• Lead with --show-comparison so stakeholders see the chunking/compression win.
• Set --region to your bucket's region — pricing varies by region.
• Reserve --real-time-pricing for final sign-off; the built-in table is fine for exploration and needs no credentials.
• Feed estimates into budgets: size a project budget from the estimate, then enforce it — see Budgets & volume quotas.

See also

• Cost management & reporting — track actual spend after uploading.
• Budgets & volume quotas — turn an estimate into an enforced limit.
• Lifecycle & storage classes — reduce storage cost over time.
• Reference: Cost, budget & alerts commands.