> ## Documentation Index > Fetch the complete documentation index at: https://resources.athenaintel.com/llms.txt > Use this file to discover all available pages before exploring further. # Automation run history > Inspect Automation versions, step attempts, decisions, outputs, and cost Run history records each Automation execution and its step attempts. Use it to understand a result, inspect an approval wait, and check completed work before starting another execution. ![Run states: queued, running, needs input, succeeded, failed or cancelled, with redrive starting a new run from the first unfinished step, and the contents of the run record.](https://mintcdn.com/athenaintelligence-e46bc9d3/6ueVSVCwufwDQRFe/images/diagrams/automation-run-status.webp?fit=max\&auto=format\&n=6ueVSVCwufwDQRFe\&q=85\&s=1b65ed9b37e25f96ee50b9c9e266f2a8) A run moves from queued to running, can wait for input, and ends succeeded, failed or cancelled. A redrive starts a new run. ## Open a run You need view access to the Automation. Cancel and redrive actions require edit access. 1. Open **Automations** and select the Automation. 2. Select **Runs**. 3. Open a run to inspect its details. 4. Follow its run ID link to open the full run inspector. The run inspector also opens from run links in notifications. Access checks still apply when you follow a direct link. ## Read the run record | Record | What you check | | - | - | | **Version** | The published version pinned to this execution. | | **Trigger** | How this execution started. | | **Inputs** | Values supplied to the run. | | **Steps** | Step status, attempts, duration, and recorded cost. | | **Output** | The run's returned result, when present. | | **Cost** | Recorded spending for the run. | Expand steps to inspect their recorded inputs, outputs, and errors. The inspector groups loop iterations, parallel branches, and error-handling scopes under their parent steps. An attempt number identifies a repeated attempt within the same run. A copied step identifies work retained from an earlier run during a redrive. Large step outputs can be stored separately. The inspector shows a preview and provides access to the stored output when available. Some run content can be withheld when you lack access to it, even when you can read the run's summary. ## Understand statuses The table lists the UI labels and the status values used in the run record. | UI label | Status value | Meaning | | - | - | - | | **Scheduled** | `scheduled` | The run is scheduled to start later. | | **Queued** | `queued` | The run is waiting to start. | | **Running** | `running` | Execution is active. | | **Needs input** | `needs_input` | The run needs input before it can continue. | | **Succeeded** | `completed` | Execution completed. Check the outcome and step results. | | **Failed** | `failed` | Execution ended with a failure. | | **Cancelled** | `canceled` | Execution was cancelled. | An approval step can be waiting while the stored run status remains `running`. The run view shows **Needs input** when it detects that approval wait. The outcome gives more detail than the status. For example, a completed run can have a skipped outcome. Read it alongside the step results. ## Inspect approvals 1. Open the waiting run. 2. Review the approval attached to the waiting step. 3. Check the requested decision, eligible approvers, and votes already recorded. 4. In the run inspector, review the approval delivery history if a reviewer did not receive a request. You can decide only when the approval names you and you remain eligible. View access to the Automation does not make you an approver. ## Retry failed work A definition can configure automatic retries for eligible steps. The step history keeps their attempts within the same run. Retry settings allow up to ten attempts, with no backoff, fixed backoff, or exponential backoff. For a failed or cancelled live run, **Redrive** starts a new run from unfinished work. It uses the source run's pinned version and inputs, even if a newer version is published. 1. Inspect the failed step and any completed external actions. 2. Select **Redrive** in the run details or inspector. 3. In the inspector, choose **From the first step that did not finish** or a specific top-level step under **From a step**. 4. Review the confirmation and select **Redrive**. 5. Open the new run and check its results. Finished steps before the restart point are copied. The restart step and later steps execute again. A nested failure restarts its top-level container, so work inside a loop, branch, or error-handling block can repeat. Dry runs cannot be redriven. A run called by another Automation must be redriven through its parent. Budget and concurrency checks also apply to redrives. > [!WARNING] > > Check the outside system before repeating a step that sends a message, writes data, or creates a record. A failed or timed-out attempt can leave an external action completed without a recorded success. Check receipts, destination records, and any idempotency key before redriving a container or choosing an earlier restart point. Starting with **Run now** creates a separate execution of the currently published version. It does not resume the failed run. ## Cancel an active run 1. Open the active run. 2. Select **Cancel run**. 3. Review the confirmation and select **Cancel run** again. Cancellation records the run and open step attempts as cancelled and sends a stop request to execution. It does not undo completed external actions. You cannot resume the cancelled run in place; start a new run or redrive eligible unfinished work. ## Related * [Automations](https://resources.athenaintel.com/docs/pillars/applications/automations.md) * [Automation approvals](https://resources.athenaintel.com/docs/pillars/applications/automation-approvals.md) * [Automation budgets](https://resources.athenaintel.com/docs/pillars/applications/automation-budgets.md) This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.