> ## Documentation Index
> Fetch the complete documentation index at: https://help.flowforth.co/llms.txt
> Use this file to discover all available pages before exploring further.

# Retrying & Resuming Runs

> Re-run a failed or canceled automation from where someone left off — or from a step you pick.

Sometimes a run doesn't finish the way you wanted: an email failed partway through, or you canceled a run you actually meant to keep. **Retry** lets you re-run an automation for those people — without rebuilding anything.

Retries always run against the automation **as it exists now**. Because they resolve against your live steps rather than a saved snapshot, even runs from before step history existed can be retried, and any steps you've since deleted simply aren't offered as starting points.

***

## Which Runs Can Be Retried

Retry is available for runs that **failed** or were **canceled** — the ones that stopped without finishing. A run that's still **In Progress** shows **Cancel Run** instead, and a run that already **Completed** doesn't offer Retry, since it finished as intended.

You'll find Retry on an automation's [**People** tab](/automations/managing/monitoring):

* **One person** — click **Retry** on their row.
* **Several at once** — select runs with the checkboxes, then click **Retry** in the action bar (or press **R**).

If your selection includes runs that can't be retried — anything that isn't failed or canceled — the tooltip and retry dialog say so, and only the eligible runs are retried while the rest are skipped. If none of the selected runs qualify, the **Retry** button is disabled and hovering it explains why.

***

## Choosing Where to Pick Up

When you retry, a dialog asks how each run should resume. There are two modes.

### Resume where each left off

Every selected run continues from its **own next step** — the one right after the step it last completed, in the automation **as it exists now**. In most cases nothing that already happened runs again. But if you've **reordered or inserted steps** since their original run, an email they already received can fall on or after their pick-up point and be sent again — the retry dialog warns you when that's a possibility.

If a run has no next step to resume from — for example, its last completed step was deleted, or it already reached the end of the flow — it's skipped. When none of your selected runs can resume this way, the option is unavailable and you'll pick a step instead.

### Start everyone from a specific step

Pick a single step, and **every** selected person runs the flow from there. The step picker mirrors your builder: top-level steps, plus each [**If / Else**](/automations/builder/if-else) condition's **Yes** and **No** paths.

* **Starting at a condition** re-evaluates it, routing each person down the Yes or No path based on their list membership at that moment.
* **Starting inside a branch** forces everyone onto that path, skipping the condition check above it. Flowforth warns you when the step you picked is inside a branch.

Before you confirm, a preview tells you how many emails will send from that point. Keep in mind that **any email a person already received will be sent again** if it falls after the step you start from.

***

## What Happens to the Run

When you retry, the failed or canceled run **becomes active again** and picks up at the step you choose — it turns back into an **In Progress** run on the People tab rather than spinning off a separate run beside the old one. You'll see the status flip as soon as you confirm.

Because it resumes at the step you choose, steps **before** that point are never re-executed. Since a resumed run continues through the automation **as it exists now**, an email someone already received can be sent again if you've reordered or inserted steps since their original run — the retry dialog flags this before you confirm. Per-step send limits work exactly as they do on a normal run.

<Note>
  Retrying is always something you start yourself — Flowforth never
  automatically retries a failed run. If a person still has a separate
  **in-progress** run for the automation, retrying one of their stopped runs is
  skipped until that run finishes.
</Note>

***

## Retry All Failed

When a lot of runs fail at once — most often because an automation hit your plan's **send limit** — you don't have to select them row by row. A **Retry all failed** button appears at the top of the automation's [**People** tab](/automations/managing/monitoring) whenever there are failed runs, and retries every failed person in a single pass.

<Warning>
  Runs that fail because you hit your send limit do **not** resume on their own.
  Once your limit resets or you raise it, use **Retry all failed** to send the
  rest.
</Warning>

Retry all failed works just like a bulk retry, with a few conveniences for doing it at scale:

* **Everyone failed is included** — no selecting rows. Each person is retried **once**, even if they have several failed runs; their latest run is the one that resumes.
* **Choose where to pick up** — the same two modes as a bulk retry: **Resume where each left off**, or **Start everyone from a specific step**. See [Choosing Where to Pick Up](#choosing-where-to-pick-up) above.
* **Optional date range** — narrow the retry to failures within a date range if you only want to re-run a specific window.
* **Live count** — the dialog shows how many failed people match your current filter, so you know exactly how many runs you're about to retry before you confirm.

***

Want to see exactly where each person got before you retry? Open their [step timeline](/automations/managing/monitoring#step-timeline) from the People tab.