Set up dashboard alerts. Know what needs attention.

Start with one workflow and one destination

Dashboard alerts can highlight a completed run, an attention state or a reported number that crosses a threshold. Choose a rule, its Simple or Warning presentation, and the destinations you want to receive it.

This October 2026 guide covers the customer dashboard. Available features depend on your licence. The tutorial uses the local dashboard interface with fictional devices and sample values; verify the setup and delivery in your own environment.

Begin with a non-critical workflow whose normal behaviour you know. Keep its State change rule and Simple style for a first test. Add a shared channel after the dashboard behaviour is clear.

• The row bell controls your personal dashboard alerts.
• Enable Alerts in the toolbar subscribes this browser or device to browser push.
• Discord and Slack have separate integration settings and switches on each row.
• Dashboard settings are separate from alerts configured in the desktop application Preferences. This dashboard has no Email, SMS or Telegram delivery controls.

1. Configure a row

Open Table and find the device or counter using search or the Source filter. Expand its row to reveal Row Settings and Alert Rules. Set Alerts Enabled to On, or enable its bell. A row with no rules receives a Simple State change rule automatically.

Keep that rule or select + Add rule, choose State change or Variable, and complete the fields. Choose Simple or Warning, then click outside the edited field and allow it to save. These row settings save automatically; they have no separate Save button.

Each rule is an independent condition: any matching rule can alert. Several rules do not form an AND condition. A row displays one active alert at a time. To remove a rule, use Remove; to stop notifications, switch off the relevant bell or channel. Removing every rule while monitoring stays enabled can restore the default State change rule.

Live Feed uses the same saved controls, but one tile can represent several counters. Selecting a partially enabled control enables all counters on that tile; selecting a fully enabled control switches them all off. Use Table for individual counter setup.

2. Understand state change alerts

A State change rule covers all four attention states below. There is no separate selector for each state. Running, Idle and update progress do not trigger it.

Offline means recent updates are missing. It does not prove the device has powered off. Check Last seen, the device application and its connection before sending commands. Monitoring intervals can make the row label and an external notification appear at different times.

Repeated updates in the same state do not notify on every refresh. After acknowledgement, another qualifying transition is needed. State alerts normally clear when the same source returns to a non-alerting state, such as Running or Idle; shared delivery may still be pending.

• Aborted: the workflow reports an aborted run. Read the reason and history, inspect the screen and correct the cause before restarting.
• No Triggers: a running workflow has exceeded its inactivity window without a new trigger. Check whether the task is waiting normally.
• Max Complete: the counter reached its configured loop target. Review the output; this can be expected completion.
• Offline: the dashboard is missing recent device updates. Check their age and restore a current view.

3. Allow for normal waiting time

Set No Triggers (min) to a whole number of minutes longer than normal pauses. For example, 10 allows ten minutes without a new trigger while Running. The standard default is five minutes, but existing rows or devices may differ.

Enter 0 to disable this inactivity check while leaving the other State change conditions available. For a native workflow, the device must receive the setting change before the new timeout applies.

Where available, Profile > Exclude from No Triggers exempts the active profile from this inactivity state. This changes monitoring behaviour. Acknowledging an existing alert does not provide the same exemption.

4. Monitor a reported numeric value

Select + Add rule, choose Variable and enter the name exactly as it appears in the reported variables. A value kept only inside the workflow cannot be monitored here. For a low-stock example, use stockRemaining, the < operator and Value 10.

Use plain numbers such as 10 or 10.5, without units or currency symbols. Single-value comparisons use Value; Between and Outside use Min and Max, with Min no higher than Max. Choose the alert style and leave the field to save.

A missing variable does not trigger the rule. Use consistently numeric readings: non-numeric values can give unexpected results, especially with not equal, so this is not a validity check for incoming data.

An alert normally fires when the condition becomes true. To test again, acknowledge it and let the dashboard observe the condition clear before it matches again. For stockRemaining < 10, test 12, then 9; acknowledge; return to 12, then 9.

• = matches Value exactly; not equal matches a different numeric value.
• > and < are strict: stockRemaining < 10 matches 9, not 10.
• Between includes both endpoints: Min 20 and Max 30 includes 20 and 30.
• Outside excludes the endpoints: the same limits match values below 20 or above 30.

5. Choose a style and respond deliberately

Simple gives a flashing bell, a highlighted row and a short sound. Warning adds an interrupting dialog with the reason and screenshot controls, plus a longer sound. Sound depends on browser permissions, tab muting and device volume; the style does not enable an external destination.

Read the reason before acting. The available screenshot may predate the alert. Capture now requests another image where supported; allow time for the response. An offline device or one without capture support may not provide a fresh image.

The dashboard alert rule Auto-stop checkbox is currently inactive in this delivery flow. Use Stop workflow or the row stop control and confirm the updated device state. For an automatic stop at a limit, configure the stop within the workflow itself.

• Close hides the Warning dialog and leaves its alert active.
• Dismiss, Dismiss Alert or a flashing bell acknowledges the current alert while keeping monitoring enabled.
• Stop workflow sends a stop request for the affected row or counter. Verify the device response.
• Dismiss current state is a separate row action for Aborted, No Triggers and Max Complete. It shows Dismissed until the condition clears; it does not stop the workflow.
• Acknowledgement does not stop or restart execution, fix the condition or remove messages already delivered.

Understand personal and shared controls

The dashboard bell and its acknowledgement belong to your login. Row rules and No Triggers settings also affect the shared row configuration used for channel notifications. Discord and Slack integrations are shared across the licence, with separate switches on each row.

Switching off a personal bell can leave shared messages running. Disable the row channel icon to pause one row, or switch off the integration master and select Save to pause that destination across the licence.

Selecting a quiet enabled bell switches personal monitoring off. If it is flashing, acknowledge first and then switch the quiet bell off, or set Alerts Enabled to Off. Dismiss alarms acknowledges personal alarms and keeps bells enabled; Clear alarms switches personal bells off. Neither disables shared Discord or Slack delivery.

Operational row alerts currently have no scheduled quiet hours or snooze. The 24-hour Snooze under Profile edits belongs to that separate review list.

6. Enable browser push on each receiving device

Sign in on the browser and device where notifications should arrive. Select Enable Alerts in the toolbar and allow notifications when asked. The button becomes Disable Alerts when the subscription is active. Enable a personal row bell and a matching rule for each row you want to monitor.

Repeat the subscription setup on every receiving browser or device. Disable Alerts removes the current subscription only; it leaves row rules, dashboard bells and shared channels unchanged.

If push is blocked, check site permission, operating-system notifications and Do Not Disturb. A missing Enable Alerts button can mean the browser does not support push. If the server reports push is not configured, contact support. Delivery while a browser or device is closed, asleep or offline depends on platform support.

7. Connect a Discord destination

In Discord, create a webhook under Server Settings > Integrations and select the receiving text channel, or ask your server administrator to do this. In the dashboard, open Discord, select Enable Discord notifications and paste its Webhook URL.

Choose Attach latest screenshot when available if needed. Select Save, then Send test, and check the intended channel. Enable the Discord icon on each row to monitor. A native row with no rules receives a default State change rule.

Use a suitable test workflow to confirm a real row event also arrives. Send test checks the saved destination, not every row configuration. An image depends on capture availability; a message can arrive without one. Choose a channel appropriate for the screen information being shared.

A blank saved Webhook URL field preserves the existing value. Enter a new URL to replace it, or use Remove webhook to clear it. To pause delivery across the licence, switch off Enable Discord notifications and select Save.

8. Connect Slack and check screenshot delivery separately

For text alerts, create or use a Slack app, enable Incoming Webhooks and authorise the receiving channel. In the dashboard Slack settings, select Enable Slack notifications, enter the Incoming webhook URL, select Save and then Send test. Check the channel and enable the Slack icon on each row to monitor.

For screenshots, ask the Slack app administrator for a bot token with message and file-upload permissions, including chat:write and files:write, and access to the destination channel. Enter the token and the Channel ID, rather than a name such as #alerts. Select Attach screenshots and Save.

Send test checks text delivery only. Trigger a suitable real alert on a device with capture support to verify image upload. A webhook alone sends text even when Attach screenshots is selected. You can save both webhook and bot settings, using the webhook as a text-only fallback.

Leave saved secret fields blank to retain their values. Use Remove webhook or Remove bot token to clear them. Switch off Enable Slack notifications and Save to pause this destination across the licence. Keep webhook URLs and bot tokens out of screenshots and support messages.

Handle other dashboard prompts in their own context

Human Check Required is a workflow request, separate from State change and Variable rules. Read the message, inspect the screenshot and review the current step. Choose Continue after completing the requested check, or Stop Automation to stop. Inspect the device first if an image is unavailable.

Where enabled, Profile edits opens Local Profile Edits for review of local copies made from shared profiles. Mark in progress tracks review, Snooze defers the notification for 24 hours, and Dismiss acknowledges it. Share profile and Delete local profile affect the profile itself; review the details before using them.

If the dashboard reports that it is not live, refresh and confirm updates resume. Application Update markers and account or trial notices are separate from configurable row alert rules.

9. Verify one complete alert cycle

1. Start with one non-critical workflow and one destination. Save the integration and confirm its test message reaches the intended channel.

2. Enable the row controls and create a real qualifying change, such as a normal test run reaching its loop target. For a numeric rule, use a reported test value that moves into the matching range.

3. Check the dashboard and destination. If screenshots matter, verify the actual image attachment and whether it reflects the event. Confirm fresh device state after any stop request.

4. Acknowledge the alert, allow the condition to clear and test a later qualifying event. Apply the verified setup to other rows, avoiding unnecessary duplicate rules.

Troubleshoot the rule, destination and response

Start with the row and the time of the expected event. Check incoming updates before adjusting thresholds or sending commands. These checks help distinguish a missing event from a delivery or display issue.

For support, include the device and counter, approximate event time, rule, destination and any visible error. Exclude webhook URLs and bot tokens from the material you share.

• No row alert: check the correct counter, Alerts Enabled, rule and fresh updates. Acknowledge any existing alert and confirm a new qualifying event occurred.
• Test message works but real alerts do not: check the row channel icon and rule separately. Confirm the event occurred after setup.
• Numeric rule never triggers: confirm the exact reported variable name, a numeric reading, operator and bounds. Let the condition clear before retesting.
• No Triggers appears too often: allow normal waiting time, or use the profile exemption where appropriate. A value of 0 disables that inactivity check.
• Push or sound is missing: check the current browser subscription, site permissions, operating-system settings, tab muting and volume.
• Discord icons are missing: save a valid webhook and enable the integration, then refresh if needed.
• Slack sends text without an image: check bot permissions, Channel ID, channel access, Attach screenshots and device capture support. Send test does not check images.
• A screenshot is missing or old: check device connectivity and capture support, and use Capture now where available.
• Dismiss did not stop execution: use Stop workflow or the row stop control and verify the response. The dashboard Auto-stop checkbox is currently inactive.
• Messages continue after the bell is off: pause the separate row channel icon or integration master, and Save the integration change.
• An alert does not repeat: acknowledge it and allow the condition to clear before a new event. Repeated identical readings are suppressed.
• The dashboard is stale: refresh after a connection warning and check Last seen, network access and the device application.