Watchers and Notifications
Watchers are screen-less automation. You write down the condition you want monitored — like "Notify me if line 3 pressure exceeds 90" — in plain language, and Studio checks the field data at a set interval, notifying you only when the condition is triggered.
From the left sidebar, go to the radar icon (fa-radar, Watcher). If there are unread notifications, a red number badge appears on the icon (10 or more shows 9+).
The screen is divided into two tabs.
| Tab | What It Does |
|---|---|
| Notifications | List of alarms triggered so far (default tab) |
| Settings | Create, enable/disable, run immediately, and delete watchers |
Creating a Watcher
Go to the [Settings] tab → New Watcher area to create one.
1. Get a Draft from a Single Sentence of Natural Language
In the input field at the top (예: 매시간 3번 라인 압력이 90 넘으면 알려줘), write your monitoring condition as a sentence and click AI Fill (fa-wand-magic-sparkles). The Name, Instructions, and Execution Interval (minutes) fields below are automatically populated. Pressing Enter works as well.
Here are common examples used in the field.
용해로 1호기 온도가 850도를 넘으면 알려줘
3번 라인 설비가 10분 넘게 정지 상태면 경고해줘
지난 1시간 CRITICAL 알람이 5건 이상 발생하면 알려줘
2공장 압축기 가동률(OEE)이 80% 밑으로 떨어지면 매시간 알려줘
보일러 급수 유량 태그가 30분 동안 값이 안 들어오면 알려줘
2. Review and Create
The auto-filled values can be edited.
- Name — The title shown on notification cards and webhook messages. Choose a name that's recognizable in the field.
- Execution Interval (minutes) — 1 to 1440 minutes. Default is 60.
- Instructions — The task description the agent will carry out each time (
예: 3번 라인 OEE를 확인해서 80% 미만이면 경고. /oee?site_id=S1 조회.). The more specific you are about which equipment, tags, and thresholds to compare, the fewer false positives you'll get. - Webhook URL (Optional · Slack/Teams) — See Exporting via Webhook below. If left blank, notifications only appear in Studio.
Click Create to register it as active immediately; it starts running automatically from the next interval.
For equipment shutdown or safety-related checks, use 5–10 minutes. For process values like temperature or pressure, use 15–30 minutes. For slowly changing metrics like OEE or production output, use 60 minutes or more. Shorter intervals increase AI usage.
3. Create Directly from Chat
You don't need to enter the Watcher screen. You can also create a Watcher by saying the following in Chat.
3번 라인 압력이 8bar 넘으면 알려줘
용해로 온도 15분마다 감시해줘
The agent first finds the relevant equipment and tags to confirm their IDs, then creates the Watcher and confirms in its reply what it's monitoring and at what interval. The newly created Watcher appears immediately in the [Settings] tab list.
Receiving Notifications
When a condition is triggered, you're notified in three places at once.
- Real-time toast — If you have Studio open, a notification in the form of
🛰 와처이름 — 내용appears at the bottom right of the screen for about 6 seconds. - Navigation badge — Unread count on the Watcher icon in the sidebar.
- [Notifications] tab — Full trigger history. Cards pile up showing the warning icon, watcher name, time of occurrence, and message.
Opening the [Notifications] tab marks all notifications up to that point as read, and the badge disappears.
You can filter by content or watcher name using the search bar above the list (알림 검색…). It shows 30 items at a time, and you can load more with Load More. If nothing has happened yet, it shows "No new notifications / Items will appear here when a watcher detects an alarm."
Exporting via Webhook
If you enter a Slack or Teams incoming webhook URL in the Webhook URL field when creating a Watcher, alerts will also be sent to that channel when triggered. The message format is 🚨 [와처이름] 내용.
Webhooks and real-time toasts are sent only once, at the moment a condition transitions from normal to alarm. While the condition persists, it won't send again on every interval. It sends a new notification only after the status returns to normal and then gets triggered again. This is designed to prevent channels from flooding overnight. (Execution history itself is recorded every interval, so you can check progress in the [Settings] tab.)
Managing Watchers
From each watcher card in the [Settings] tab, you can do the following:
- Enable/Disable — Toggle on the left side of the card. Turning it off stops scheduled execution; turning it back on resumes it. If you're in the middle of an inspection, turn it off rather than deleting it.
- Run Now — Executes once immediately, without waiting for the interval. Use this to verify that a newly created watcher's instructions work correctly.
- Delete — Trash icon (
fa-trash-can). Deleted after confirmation and cannot be undone. - Execution History (n) — Expanding shows recent runs with time and a colored result dot.
Each card displays the execution interval ({n}분마다), Last Run time (if it has never run, 실행 전), and the last result badge.
| Badge | Meaning |
|---|---|
| Normal | Ran and condition was not met |
| Warning | Condition met — notification sent |
| ERR | Run failed (data query error, AI not configured, etc.) — reason shown in the body below |
The name, instructions, and interval of an existing watcher cannot be edited directly in the UI. To change a condition, create a new watcher and delete the old one. If you only need to pause it, simply turn off the toggle.
If the bottom status bar shows "Platform Disconnected — Real data and watchers temporarily paused" , watcher executions are skipped. The execution history will note "Execution skipped due to platform disconnection (auto-resumes upon reconnect)" and will automatically resume from the next interval once the connection is restored. This is intended to prevent false alarms. If the connection stays down for a long time, check Troubleshooting.
Permissions
- Admin / Builder — Can create, enable/disable, and delete watchers.
- Viewer — Can view the [Notifications] tab and the watcher list, but the creation form and the run/delete buttons are not visible.
Watchers and notifications can also be checked and managed from your field phone — see Field Mobile.