Skip to main content

Test, Debug & Publish Workflows

Designing a workflow is only half the process. To ensure reliability and trace correctness, StreamSpace Studio provides built-in testing tools, execution logs, and version control options. This guide walks you through verifying your logic before deployment.

1. Running Tests with "Run Once"​

The Run Once tool executes your workflow directly on the canvas using mock inputs, avoiding production impacts.

Triggering a Test Run​

  1. With your workflow open on the canvas, click the Run Once button in the top-right toolbar.
  2. If the debug/preview panel is open, you can monitor execution live under the PREVIEW section.
  3. The canvas visually traces the active execution path, highlighting each node with blue or green borders and checkmarks as control passes through.
Visual Path Tracing

Watch the canvas execution path carefully: green outlines and checkmarks indicate successful node runs, while red outlines indicate logic splits or error branches.

2. Debugging & Subscriber Logs​

If a test run encounters errors, or output variables are not what you expected, inspect live node metrics and data payloads via the Debug panel.

Opening the Debug Panel​

  1. Click the Debug button in the top-right toolbar.
  2. In the View Debug - Select Flows dialog, select the workflow flow name or flow ID you want to inspect, then click Confirm.
  3. This opens a split view containing the live execution PREVIEW on the left and Subscriber Logs on the right.

Analyzing Logs, Payloads, and Execution Nodes​

  1. Subscriber Logs Drawer:
    • View chronological session entries containing timestamps, session IDs, workflow names, and executed node names (e.g., Manual Trigger, Get Sheet, Core Agent, Send Text Message).
    • Click the expand arrow on any log entry to view full JSON execution payloads (including inputs, outputs, memory, status, and metadata).
    • Click "Jump to this node" next to any log entry to immediately pan the canvas to that specific node.
Canvas Highlighting

Clicking the "Jump to this node" link in any Subscriber Log entry immediately centers and zooms the playground canvas onto that specific block, saving navigation time on complex topologies.

  1. Preview Drawer:
    • Displays real-time step-by-step workflow outputs, human-readable text generation, tool responses, and node execution statuses (showing runtime duration, e.g., 1864 ms or 21485 ms).
    • Expand any individual node entry to review its detailed INPUT, PROCESS DATA, and OUTPUT structures.
  2. Click the Clear (trash icon) or Close (X) button at the top of the Subscriber Logs panel when finished debugging.

3. Pre-Publish Checklist​

Before pushing a workflow to production, verify the following configuration points to prevent live errors:

  • Check Node Connections: Ensure there are no loose nodes on the canvas. Every node (except terminating nodes) should have wired output arrows.
  • Validate Required Fields: Ensure all required parameters in node settings are fully defined (e.g., prompting values, message content).
  • Authenticate Connectors: Verify that Google Sheets, Email/Messaging, Database, or Webhook nodes are using valid, authorized credentials.
  • Define Error Paths: Connect error outputs (🔴 ports on logic or integration nodes) to fallback nodes to catch failed calls gracefully.

4. Publishing & Version Control​

Publishing freezes your current workflow state into an immutable release version. Published versions can be assigned to AI Assistants, Agentic AI Reports, or managed across historical snapshots.

Publishing a Version​

  1. Click the Publish button in the top-right toolbar.
  2. In the Name this version modal:
    • Name *: Enter a semantic version tag (e.g., 1.0.0, 1.0.1).
    • Release Note: Document what changes or fixes were introduced (e.g., "Changed connection", "Initial setup"). Maximum 200 characters.
  3. Click Publish. A toast confirmation ("Flow Created Successfully") will confirm the release.

Managing Workflow Versions​

  1. Click the Versions icon (located in the top-right menu next to Publish).
  2. The Versions side panel lists all published releases with their version numbers, timestamps, author details, and release notes.
  3. Restoring Previous Releases: Select an older version from the list and click Restore at the top of the panel to revert canvas settings to that specific release.
  4. Click Exit Versions to return to regular canvas editing mode.
Production Deployments

Only published versions with tag names (e.g., 1.0.0) can be mapped directly to AI Assistants or Agentic AI Reports. Re-publishing a flow with the same release number is not permitted.