Skip to main content
Go to the Studio
Tutorials

Running and monitoring actions

Running and monitoring actions

While a robot runs, it matters that you understand how its actions are executed and how to follow their progress. This guide covers the run options and how to watch an action's status live.

Run options​

Run options sit at the bottom of every action's settings panel and control what the robot does if the action runs into trouble.

Stop the robot on error​

Whether the robot stops when this action fails.

  • Yes (the default): the robot stops entirely and the run ends in an error
  • No: the robot skips the failed action and carries on with the next ones

When "No" makes sense:

  • Optional actions that are not critical to the flow
  • When you want the robot to push on despite a partial failure
  • In validations where you want the error recorded but not the process stopped
caution

Use "do not stop on error" carefully. If later actions depend on the result of the one that failed, they may fail too, or produce something unexpected.

Number of attempts​

How many times the action is retried before it counts as failed.

  • Minimum: 1 (the default)
  • Recommended: between 1 and 3, depending on the action

When more attempts help:

  • Actions that depend on the network (downloading files, HTTP requests)
  • Interactions with elements that may take a moment to appear
  • Operations that can fail briefly because something is unstable

For example: set to "3 attempts", an action that fails twice and works on the third try counts as a success.

The visual states during a run​

Running the robot in test mode updates each action's card as it goes, so you can watch what the robot is doing in real time.

StateLookWhat it means
RunningA blue border, pulsingThe action is executing right now
SucceededA green badge with a checkThe action ran with no errors
ErrorA red border with an error badgeThe action failed after all its attempts
SkippedFadedThe action was skipped because an earlier one failed (in "do not stop" mode)

Looking at an error​

When an action fails, click its card to open the settings panel. There you get:

  • The error message: what went wrong
  • Technical details: extra information for diagnosis
  • Attempts made: how many times the action ran before failing

That is usually enough to work out the problem and fix the flow.

The full cycle​

The complete cycle of using an action, from configuration to monitoring:

1. Open the catalog
   ↓
2. Pick the action you want
   ↓
3. Configure it in the panel:
   ├── Set the action's ID (e.g. "checkBalance")
   ├── Pick the mode (where there is one)
   ├── Fill in the input fields
   ├── Enable the output fields you will use
   └── Set the run options
   ↓
4. Save
   ↓
5. Next action: reference this one's outputs
   ↓
6. Run the robot
   ↓
7. Watch the cards' status
   ↓
8. On an error: read the message in the panel

Good practice​

Set the attempts sensibly​

Do not pile up attempts on an action that is unlikely to work however many times it repeats. It only makes the run slower.

Use "do not stop" with care​

Always ask whether the following actions can work without the result of the one that might fail. Where there is a dependency, it is better to let the robot stop.

Watch the first run​

When you are testing a new flow, follow the run visually — it is the fastest way to see where the problems are.

Write down the common errors​

If an action keeps failing, note the cause and the fix. It makes the next round of maintenance much shorter.

Next​