Skip to main content

Production Batch

Prototype with priority 0. Use priority 1 for the one-hour production SLA and higher quotas. Production jobs normally execute a published Function so the behavior you evaluated is the behavior applied to the full dataset.

Machine-time pricing

Successful Batch jobs are billed for recorded GPU inference time. The default rate is $3.00 per GPU-hour, saved when each job is created.
For example, two GPUs running inference for 30 minutes use one GPU-hour and cost $3.00 at the default rate. Startup, queueing, and failed retries are excluded. Charges are rounded to four decimal places. Estimates, samples, and restarts retain the job’s saved rate. Later rate changes apply to new jobs. See the job’s machine_time_pricing fields for its measured GPU hours, hourly rate, and charge in the job response.

Estimate before submitting

A dry run creates an estimate job and returns its job ID. It does not launch the full workload.

Run a Function on S3 with recovery

For a published Function and one S3 CSV, Parquet, or JSONL object, the CLI manages signing, a durable receipt, progress watching and normalized Parquet download:
Configure deployment credentials and local AWS access. Repeat the same command and output path after interruption, keeping results.parquet.sutro.json and partial-download files. With a known job ID, sutro jobs resume JOB_ID -o recovered.parquet watches/downloads without source AWS credentials. A fresh output path on batch-run can create another submission. In Python, use batch_run_function() with a signed S3 URL, persisted idempotency key and pinned source identity. It returns a job ID after input acceptance; use the existing job status and results/download methods afterward. Python calls do not create file receipts. Both interfaces require a deployment with resumable submission support. Timeout and Ctrl-C leave remote work running. A deliberate new run can incur new compute charges; overwriting a local result does not itself submit work.

Submit an HTTPS input

For large inputs, pass a short-lived HTTPS GET URL for a CSV or Parquet object. We recommend using a presigned URL with a 1-hour expiration, as Sutro needs to be able to download the file within that time frame.
Use an immutable object key and a read-only signing identity. Treat the full presigned URL as a temporary credential.
CSV or Parquet columns must match the published Function’s inputs. See your Function’s configuration for the required columns.

Wait and download

Avoid materializing a large result as JSON:
Parquet downloads support resuming after interruption. Keep your source object until you have reconciled the result rows.