# Runs

Congratulations on having a workflow that's ready to run! Here's some basic information and suggestions on starting, viewing, and managing your runs.

### Test runs

If you're validating your workflow, click `Start a test run` in the `Editor` tab to run your workflow all the way through. Test runs are the same as real runs, but you'll select existing data to use (e.g. an existing row in a spreadsheet) rather than waiting for a trigger.

**You have a reasonable number of free test runs**, and they reset every week! [Learn more about testing.](/running-workflows/testing.md)

You'll see a summary of how many AI credits a step would have used in a non-test run at the top of the run once it's complete.

You can only have one test run at a time. When you make changes to a workflow with an active test run, you'll be prompted to update or replace the existing test run.

<figure><img src="/files/TVJV112yjjqCvppvDD7z" alt=""><figcaption><p>Test runs (as well as step tests) don't cost AI credits. You'll see what it would have cost as a real run at the top of the run.</p></figcaption></figure>

### Runs tab

To see your workflow's active runs, click on the `Runs` tab at the top of the screen. Each run is a specific instance of your workflow.

<figure><img src="/files/l937PAZJE6LmGkyOE41l" alt=""><figcaption><p>Runs tab</p></figcaption></figure>

Runs can be in several states. You'll see these states grouped in sections in the `All runs` list in the left navigation.

* **Pending**: Runs that haven't started yet, like a large batch still processing
* **Running now**: Runs that are actively working on the latest step
* **Issues**: Runs that are paused due to errors
* **Action required**: Runs that require a human-in-the-loop action, like entering data or approving that it can continue
* **Waiting**: Runs in a waiting state, for example due to a wait step or being manually paused
* **Completed**: Runs that have fully completed
* **Trash**: Deleted runs, which can be restored for 30 days before being permanently deleted. Runs older than your workspace's data policy (typically 30 days) will also move here.

<figure><img src="/files/0QOOrIhBj29xOa2CTaoI" alt=""><figcaption><p>Run view with some runs in an error state, waiting for human action, and completed.</p></figcaption></figure>

### Run actions

You can take different actions on runs based on their state:

* **Pause**: Stop all activity on a run until it's manually unpaused
* **Retry failed steps**: Rerun steps with errors to see if any recent updates to the workflow, or transient errors from the integration, have been resolved
* **Replay**: Start a new run using the same data as the selected run
* **End**: Make the run end right away, even if not all steps have completed
* **Delete**: They'll be in the trash for 30 days

Some sections will let you take these actions in bulk. For example, you can `Retry all failed steps` for all runs in the `Issues` section.

### Freezing runs

{% hint style="info" %}
Wait, stop! Something going sideways and you need everything to stop at once? Free all runs!
{% endhint %}

If you need to immediately stop the activity of a workflow, say because you have many active runs and you realize they all need to be stopped, look for the `Freeze all runs` button in the top right menu of the workflow or any run.

Freezing runs means:

* No new runs will be generated, even if your trigger conditions are matched
* All runs will immediately stop processing

While runs are frozen, you'll be able to delete them. When you unfreeze, any active runs will continue again, and if there's a queue of missed trigger items, new runs will be created for them.

<figure><img src="/files/uL4K3iWNeDxP7v5iOhF8" alt="" width="473"><figcaption><p>Freeze all runs: Your workflow's emergency "Wait, stop!" button</p></figcaption></figure>


---

# Agent Instructions: Querying This Documentation

If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://docs.relay.app/running-workflows/runs.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.
