본문으로 건너뛰기
This page is not translated yet. 한국어 is switched on so the navigation and URLs can be tested; the text below is the English original. Use the language selector in the header to switch back, or see the documentation home.

Automations

how-to · Reviewed 2026-07-31 GA

Scripts attached to a job that run at a defined point in its lifecycle. The most common use is extracting metadata from solver output so nobody has to enter it by hand. They also handle notifications, format conversion, and file movement.

Two different things are called "automation"

Rescale Automations (this page) are your scripts, running around a job.

Intelligent Resource Automation is a platform-managed set of features that pick hardware and balance load for you. It lives under Platform. Different capability, confusingly similar name.

Rescale Automations​

Rescale Automations let you attach scripts to a job that run at a defined point in its lifecycle. Use them to extract metadata, convert output formats, trigger notifications, move files to another system, or chain follow-on jobs, without adding that logic to your analysis submit script.

Execution timing and supported languages​

An automation can run at one of three points:

  • Before job execution, to prepare inputs or validate settings.
  • During job execution, typically as a side-car that reads files as they are produced.
  • After job completion, to post-process results.

Scripts can be written in Python or Bash.

Common use cases​

  • Metadata extraction. Capture key simulation metrics (convergence stats, residuals, runtime) automatically, and write them into custom fields or tags. This is the automation type that makes metadata capture reliable — a value nobody has to remember to enter is a value that's actually there.
  • AI extraction. Convert simulation results to formats compatible with Rescale AI Physics, such as PyVista, for surrogate model training.
  • Notifications. Post job completion messages to Slack or Microsoft Teams.
  • File management. Compress results, convert formats, or copy artifacts to cloud storage.
  • Data visualization. Render quick-look plots as part of post-processing.

Creating and using an automation​

In the UI, during job setup:

  1. Go to the Automations section.
  2. Click New Automation.
  3. Provide a name, description, execution timing, the script itself, and any optional variables the script needs.

When creating a new job:

  1. Open the Automations page.
  2. Select one or more automations to attach.
  3. Provide values for any required variables.
  4. Submit the job.

Automations run with access to the job's working directory and can call the Rescale API using the embedded RESCALE_API_KEY environment variable, which is short-lived and expires when the job completes (see embedded per-job keys).

Writing the script

Writing an automation covers the author's side: the runtime the script gets, taking configuration as variables instead of hardcoding it, the API operations automations usually call, and when to reach for a submit script or a workflow instead.

Platform-managed hardware selection​

Intelligent Resource Automation is a separate, platform-managed capability: License Queuing, Region Load Balancing and Coretype Sets, plus the Coretype Collections and Compute Optimizer recommendation features. None of it is scripted by you, which is why it lives under Platform rather than here.

See Intelligent Resource Automation and Hardware.

  • Writing an automation — the script runtime, variables, and the operations to call.
  • REST API for driving automations and reading job state from scripts.
  • CLI for chaining jobs and configuring license queuing from submit scripts.
  • Tags and custom fields — what metadata-extraction automations populate.
  • Running jobs for the user-facing view of hardware selection and priorities.
  • Licenses for license proxies and Rescale-hosted licenses.