1Open the Test panel

Step 1.1

Click TEST in the right activity bar

The right activity bar has a TEST tab. Click it (or use the toolbar Test icon) to open the test panel. The panel has three modes — Workflow, Node, and Pin — selectable from a small tab strip at the top.

Step 1.2

What's the difference vs RUN?

  • TEST is for in-canvas, dev-time runs while you're authoring. Outputs land in the test panel; nothing goes to the production audit trail.
  • RUN is the real Workflow Runner — it's the same execution path callers will use, runs are recorded, and notifications fire as configured.

2Run to a target node

Step 2.1

Right-click any node → Run to this node

The workflow runs from Start, taking the actual edges and conditions, until control reaches the chosen node. The chosen node becomes the stopping point — its inputs are visible but it does not execute.

Use this to confirm that upstream branching produces the right inputs for a downstream node before you trust it to run.

Step 2.2

Provide test inputs

The test panel shows an Inputs JSON editor pre-filled with the workflow's declared input schema. Edit the values for this run; they don't change the saved workflow.

📌
Pinned outputs are honoured

If any upstream node has a pinned output, the test will use the pin instead of re-calling the system. This makes "run to node" deterministic and fast.

3Isolated node test

Step 3.1

Right-click → Test node

Test node executes only the selected node, with inputs you provide directly. Nothing upstream runs. This is the fastest way to debug a single system call or expression.

Useful when:

  • You're not sure if a connector is configured correctly.
  • You want to verify a parameter expression compiles.
  • You want to exercise a connector with a specific payload that's hard to reach via the normal flow.
Step 3.2

Read the result

The output appears in the test panel as a JSON tree, with timing and status. Errors show with the connector's response body so you can see what the upstream system returned.

4Pin a node output

A pin is a saved input or output for a single node. When pinned, the node returns the saved value during tests instead of calling the real system. Pins are kept per workflow and survive reloads.

Step 4.1

Manual pin from a result

After a test or run, right-click the node and choose Pin output. The current output is saved as the pin. The node now shows a 📌 badge.

📌 create_user · output pinned · 1 entry · last updated 2m ago
Step 4.2

Record mode

Open the PINS panel and toggle Record mode on. Run the workflow once; every node's output is captured automatically as a pin. Turn record off and you have a full set of replayable pins for offline development.

Step 4.3

Manage input pins

Right-click a node → Manage input pins. You can pin specific input values too — handy when one parameter changes per run but the rest are stable.

Step 4.4

Unpin to use real values

Open the PINS panel, find the node row, and click Unpin. Or right-click the node → Unpin.

⚠️
Pins only affect tests, not runs

The Workflow Runner ignores pins. Pins are a development-time convenience — production runs always call the real systems.

5The Workflow Runner

Step 5.1

Open it

Three ways:

  • ▶ Execute on a workflow card in the gallery.
  • Execute button on the canvas top toolbar.
  • RUN tab on the right activity bar (opens the runner inline).
Step 5.2

Provide inputs and start

▶ Run · New-hire access request
● running
INPUTS
{
  "first_name": "Jane",
  "last_name":  "Doe",
  "department": "Engineering",
  "manager_id": "u_3a8b91"
}
▶ Start run opens the live execution overlay
Step 5.3

Closing the runner doesn't stop the run

Once started, the run executes server-side. Closing the modal only hides the live view — the workflow keeps running. Re-open the runner from the same workflow to attach again.

6Watching the execution overlay

Step 6.1

Live status on the canvas

While a run is in progress, the canvas overlays each node with its real-time status: queued, running, success, failed, skipped, waiting-for-input. Edges fade or pulse to show the path being taken.

Step 6.2

Click a node to see its actual data

Click any executed node and the right panel shows the inputs and outputs of this run. The DATA panel switches automatically to last-run mode.

Workflow Runner · Liverun_a8f31…
Start0mssuccess
EntraID · Get user412mssuccess
EntraID · Add to group298mssuccess
Approval · Managerwaiting
Step 6.3

Toggle the last-run overlay

The floating-actions stack on the bottom-right of the canvas has a last-run overlay button. Toggle it off to see the clean graph; toggle it on to see status colours from the most recent run, even after it finished.

7Completing a paused human step

Step 7.1

What pauses look like

When the engine reaches a User Interaction node, the run pauses with status waiting-for-input. The node turns amber on the canvas; the timeline shows ⏸. The runner header switches to ● paused — awaiting input.

Step 7.2

Open the pending form

The right panel shows the Pending Tasks view with one row per outstanding human step. Click the row to open the form; fill the fields; submit.

📨
Forms are also delivered out-of-band

Configured channels (email, Slack, Teams) receive a notification with a link to the form. The recipient does not need to be inside Orchestration Studio to complete the step.

8Resume after input

Step 8.1

Submit advances the run

Once the form is submitted, the engine resumes from the User Interaction node. The submission becomes the node's output and downstream nodes start.

Step 8.2

If you cancel

The form's Cancel ends the run with status cancelled. The timeline marks the node skipped; downstream nodes never run. This is recorded in the audit trail.

Step 8.3

Long-running runs

A run can stay paused for hours or days. Close the runner; the workflow remains queued. Re-open from the gallery's Runs tab (or the workflow card menu) to resume monitoring.

9Console / Problems / Timeline

The bottom IDE panel has three tabs that surface what happened during a run.

TabContents
ConsoleStreaming log lines from the engine and any Log nodes you used. Filter by level (info / warn / error). Use the Clear button between runs.
ProblemsValidation issues that need fixing before the workflow can execute cleanly: missing required parameters, dangling edges, expression errors. Clicking a row jumps to the offending node or edge.
TimelinePer-node duration breakdown for the most recent run. Useful to identify slow nodes; hover for exact start/end timestamps.
⌨️
Toggle the bottom panel

Press Ctrl + ` to show / hide the bottom panel, the same way an IDE does.