Skip to content
Adagio

Running from the UI

Use the UI when you want Adagio to manage uploads, compute-account selection, and run monitoring.

Starting a run

From the home screen, click Run next to a saved pipeline. You can also run directly from the canvas by clicking Run in the toolbar.

The run flow has two steps.

Step 1: Run arguments

The first step shows:

  • Required inputs: files or data that must be provided before the run can start
  • Open parameters: any parameters that were left configurable when the pipeline was built

Defaults on promoted parameters are prefilled. Promoted parameters with no default are required.

Providing inputs

For each required input, upload the file the pipeline expects.

Common cases include:

  • QIIME 2 artifacts such as .qza
  • metadata tables such as .tsv or .csv

The semantic type of the input tells you what shape of data is expected.

Exporting an arguments JSON template

From the run arguments step you can export a JSON template for CLI use.

Important: this export records the current input file names, not absolute local paths or Adagio upload tokens. Replace those file names with real local paths before running the pipeline with adagio run.

Step 2: Run configuration

The second step captures the UI-specific execution settings:

  • analysis name
  • analysis description
  • CPU
  • RAM
  • disk
  • account

These settings are for the Adagio-managed run. They are not part of the exported .adg pipeline file.

Submitting the run

After both steps are complete, submit the run. Adagio uploads the selected files, creates the analysis record, and starts execution against the selected account.

Monitoring progress

The run view shows each pipeline step with a status indicator:

StatusMeaning
PendingWaiting to start
RunningCurrently executing
CompleteFinished successfully
FailedEncountered an error

Click any step to see its log output. If a step fails, the log will include the error message to help diagnose the problem.

Results

When the run completes, results are available in the run view. You can download individual artifacts from there.

Reruns and caching

If you rerun a pipeline with the same inputs but different parameters, only the steps affected by the change re-execute. Upstream steps whose inputs have not changed are replayed from cache automatically.

To force a full rerun from scratch, choose Rerun (no cache) from the run menu.

Downloading for CLI use

For CLI execution, export:

  • the pipeline .adg file from the editor
  • an arguments JSON template from the run arguments step

Then use Running with the CLI and Runtime Configuration to run the same workflow locally.