# Dashboard and runs

> The automation's Dashboard tells you how many times it started, how it went and what it's waiting for. From the Run log you open the step-by-step story of each run.

Page: https://supportfast.ai/en/guide/automations/executions/ · Section: Automations · Updated on 30 October 2026

Every time the trigger event occurs, a **run** (an execution) is created. The Dashboard collects them all.

## Opening the Dashboard

In the workspace Overview press Open on the automation's row, or go to Automations › Dashboard.

At the top, choose the period from the Quick periods, or with Custom range.... The default is Last 30 days, the maximum one year. Dates follow the workspace time zone.

If the automation has never started, you read “This automation hasn't run yet” and the Open the Playground button.

## The three cards

- **Runs in the period**: How many times the automation started, compared with the previous period. Below: Completed, Failed, Stopped. A run counts in the period it was created in, even if it lasts for days.

- **Success rate**: The share of runs that ended well, again compared with the previous period. Below: Error rate and Average duration, without the waits.

- **Waiting now**: The runs open right now, whatever the period. It says “In progress” or “None in progress” and shows the next wake-up. Below: Scheduled (held on a wait) and Retrying (waiting for a new attempt or a call back).

## The charts

**Runs over time** shows the runs started in each interval, as bars stacked by status. With Granularity you choose Hour, Day, Week or Month, among those that suit the length of the period.

**Executed nodes** shows which node types worked in the period, with the total in the center.

## The Run log

Below the charts is the Run log, with the number of runs in the period. At the top are the runs still open, in order of wake-up; below, the finished ones, most recent first.

- **Filter**: All, In progress, Completed, Failed, Stopped. In progress isn't a status: it groups the runs that are still open, that is Running, Scheduled and Retrying. Remove the filter goes back to all of them.
- **Columns**: TIME, STATUS (one of the six statuses below), DURATION and MESSAGE, that is what's happening or why it ended.
- The arrow at the end of the row, Open the run detail, opens the step-by-step story.

## The six run statuses

These are the statuses of the single run, different from those of the automation as a whole (Draft, Active, Paused), explained in [Automation statuses](https://supportfast.ai/en/guide/automations/publish.md#automation-statuses). Watch out for one name: an Active automation shows as “Running” at the top, but here Running means that a single run is working right now.

| Status | When | Message in the row |
| --- | --- | --- |
| Running | The run is working right now. | “—” |
| Scheduled | It's held on a Wait node and will resume later. | “Paused until…” and the time it resumes. Only this run is waiting: the automation stays active. |
| Retrying | A node failed and will be tried again, or Make Call will call again. It also shows as Retrying while Make Call waits for a call to end. | For a call back, for example “Attempt 2/3 · next Today · 15:00”. While it waits for the call to end, the time shown is the 60-minute limit. For an error, the error sentence and “retrying shortly”. |
| Completed | It reached the end without errors. | “Completed successfully” |
| Failed | It closed because of an error: a permanent error, or the attempts ran out. | The error sentence, with the node name, in red. |
| Stopped | It was stopped without a fault in the flow. For example the contact disappeared while the run was still using it (see the frequently asked questions at the bottom), the plan didn't let it start, or the automation was paused, or had changed, when the time came to resume: see [Pausing](https://supportfast.ai/en/guide/automations/publish.md#pausing). | The reason, for example “Monthly executions used up: the automation did not run.” or “Execution cancelled: the automation was turned off.”. |

## The Execution detail

The panel opens on the right. At the top are the start, the status and the links Contact card, Open conversation or Open call. Below, the story shows node by node what happened, starting from the data of the **Starting event**: what each node sent, to whom and with which variables, any errors, and at the bottom where the run stopped or what it's waiting for.

At the bottom, Copy offers Copy JSON, the full run detail in JSON format to paste into a support request or into an AI assistant, and Copy execution id.

![Detail of a failed run: Starting event Call ended, Condition succeeded on the Appointment branch, Send Email node with the No valid recipient error and the Execution end card.](https://supportfast.ai/_astro/05-esecuzione-fallita.iiWllJ8e.webp)

_The detail shows node by node where the run stopped and why._

> **Note**
>
> **Confidential data is hidden.** Keys, tokens and called numbers appear masked, also in the text copied with Copy JSON.

## The most common errors

The error sentence appears in the MESSAGE column and in the node's error block. These are the most frequent cases.

| What happened | What to do |
| --- | --- |
| Contact Management didn't find the contact with the email or phone given. | Check the variables in the identification fields. To create the contact if it's missing, use Update contact. |
| After the variables were replaced, the email no longer has a valid recipient. | Look at Discarded recipients in the node and check the variables. |
| The recipient of an email, a template or a call is a blocked contact. | Unblock them only if it's right to do so: see [Blocking contacts](https://supportfast.ai/en/guide/contacts/blocking.md). |
| The WhatsApp template wasn't sent: empty number, template no longer approved, integration not active. | Check the number variable and the integration; if needed, choose an approved template and publish again. |
| The external service of the API node answered with an error, or the address isn't allowed. | Read HTTP status and Response in the node's card. See [Calling an external API](https://supportfast.ai/en/guide/developers/api-calls.md). |
| A plan limit: all lines busy, the month's minutes or runs used up, a node feature no longer included. | For lines, spread the calls over time, for example with a Wait. For the rest see [The run quota](https://supportfast.ai/en/guide/automations/executions.md#the-run-quota) and [Add-ons and automatic purchase](https://supportfast.ai/en/guide/account/add-ons.md). |

For an error you can't find here, write to us and paste Copy JSON, or at least Copy execution id, as explained in [How to report a problem](https://supportfast.ai/en/guide/help/report-a-problem.md).

## Retries

A node can fail for a temporary reason, such as an external service that doesn't respond. The system tries it again from that node, not from the start, a few minutes apart: there are 3 attempts in all. At the third error the run closes as Failed and the following nodes don't run.

Permanent errors close the run right away, without retries: for example an address that isn't allowed, a template that isn't approved, a request rejected by an API. The call backs of Make Call follow instead the attempts and interval set in the node.

## The run quota

Every event that starts an automation uses one run from the month's quota. In Account › Subscription you see how many you've used; how many your plan includes is in [Plans, quotas and usage](https://supportfast.ai/en/guide/account/plans-and-usage.md).

These don't use runs:

- resuming after a wait;
- retries after an error;
- the call backs of Make Call;
- restarts from the beginning for the same event, explained in [Many events at once](https://supportfast.ai/en/guide/automations/executions.md#many-events-at-once).

**With automatic purchase turned on**, the pack is bought by itself when what's left of the month's runs drops below 250. If they run out sooner, for example with many events at the same moment, the system tries to buy it at that moment. It's an automatic purchase pack (“Run pack” in the console): it's valid until the end of the month and doesn't renew. See [Automatic purchase](https://supportfast.ai/en/guide/account/add-ons.md#automatic-purchase).

**If automatic purchase is off, or the payment fails**, once the quota is used up the event is discarded. In the Run log a run appears that is already closed, Stopped, with “Monthly executions used up: the automation did not run.”. No node runs.

## Many events at once

When many events arrive, there are two different cases.

**Different events: the runs queue up.** Each event creates its own run, which uses one run from the quota. The runs of the account work a few at a time. Those for the same contact or the same conversation never work at the same moment, even if they belong to different automations. The others wait their turn and start with a small delay. For example, if you add 200 contacts from the contact book at once to the list of an automation that starts with Contact in List, 200 runs are created, which work a few at a time. To write to or call many contacts at the same moment, use a [campaign](https://supportfast.ai/en/guide/campaigns/overview.md).

**The same event, with a run still open: the run starts over.** In the same automation, if the same event arrives again while its run is still open, for example held on a Wait, a second one isn't created. The open one restarts from the beginning, with the new data, and doesn't use another run. What “the same event” means depends on the trigger:

| Trigger | The open run starts over if the event arrives for… |
| --- | --- |
| Contact events | The same contact. With Contact in List, the same contact in the same list. |
| Conversations events | The same conversation. With Tool used, also the same tool. For On every message see [Start events](https://supportfast.ai/en/guide/automations/triggers.md#on-every-message-starts-from-the-agents-replies). |
| Telephony events | Never: each call is a new conversation and has its own run, even if the same customer calls back. |
| Webhook | An identical request, with the same data. |

> **Warning**
>
> **A long wait can start over.** With Contact Updated followed by a 2-day Wait, every change to the same contact restarts the wait. With Call ended, instead, a second call from the same customer creates a second run, and the first one carries on: see example [E. Reminder the day before the appointment](https://supportfast.ai/en/guide/automations/examples.md#e-reminder-the-day-before-the-appointment).

## Frequently asked questions

**A run has disappeared from the Run log.**

A run that is still open leaves the Log, with its detail, if you delete the contact or conversation it started from, or the list of an automation that starts with Contact in List, or if you move the automation to another workspace. It doesn't stay as Stopped: it disappears.

The Stopped status with “Contact deleted. Trigger aborted.” appears in a different case: when the contact disappears while the run is still using it. For example, a Contact Management node with “Delete contact” deletes it and, after a wait, the flow tries to act on it again.

## See also

- [Saving, publishing, fixing](https://supportfast.ai/en/guide/automations/publish.md): Pausing, resuming and editing an active automation.

- [Nodes: messages and actions](https://supportfast.ai/en/guide/automations/action-nodes.md): What each node shows in the run detail.

- [Campaigns and automations](https://supportfast.ai/en/guide/help/campaign-automation-issues.md): The automation doesn't start, a message doesn't arrive.