Live Events
Real-time stream of incoming industrial machine events. Direct monitoring during critical shifts or equipment commissioning.
Live Events
The Live Events view shows the stream of industrial events in real time, at the exact moment they arrive in the system. It is like having a dedicated monitor showing everything happening in the plant right now — no waiting, no page refreshes.
What is it for?
There are situations where the history is not enough — you need to see what is happening at this instant:
- During new equipment commissioning: verify that events are arriving correctly before leaving the system running
- On high-risk shifts: plant startups, major maintenance activities, process changes
- When diagnosing a new connection: confirm that the sensor or PLC is sending data correctly
- During active incidents: see in real time how the system is responding to a critical event
The live event monitor eliminates uncertainty: you know exactly what is arriving, when, and with what information.
How does it work?
The view connects to the system through a WebSocket — a permanent connection that delivers events as soon as they arrive, without polling or delays. Each event appears on screen in under a second from when it was received by Rela AI.
If the WebSocket connection fails (due to unstable network), the system automatically switches to update mode every 3 seconds as a fallback. The feed keeps the last 200 events in memory.
How to use it?
Access from Alarms > Live in the sidebar.
Monitor controls
| Control | What it does |
|---|---|
| Live / Pause button | The pulsing green dot means you are receiving events in real time. Click to pause the feed when you want to review an event without it scrolling away. |
| Source filter | Shows only events from a specific source. Useful when monitoring specific equipment during a diagnosis. |
| Auto-scroll | Enables or disables automatic scrolling. With auto-scroll on, the view always shows the most recent event at the bottom. Disable it to review a previous event without losing focus. |
| Clear feed | Removes all events from the current view (does not delete the history). Useful to start a clean monitoring session. |
| Counter | Shows how many events have arrived since you opened the view. |
Information in each event
Each event in the feed shows:
| Field | Description |
|---|---|
| Time | Timestamp in HH:MM:SS format — updated in real time |
| Severity | Color-coded badge: Info (gray), Warning (yellow), Critical (orange), Emergency (red) |
| Source ID | Which sensor or PLC the event came from |
| Event type | Event classification (e.g., HIGH_VIBRATION, TEMP_LIMIT, EMERGENCY_STOP) |
| Status | How it was processed: accepted, processed, duplicate, throttled, etc. |
| Message | Event description |
| AI response | A 2-line summary of what the agent decided to do |
Events with Critical or Emergency severity are highlighted with a red side border — immediately visible even when the feed is moving fast.
Key benefits
- Real-time monitoring without page refreshes — events in under 1 second
- Source filter to focus on specific equipment during diagnosis
- Pause the feed to review a specific event without losing context
- Automatic fallback if the real-time connection fails
- Immediate visual identification of critical events by color and red border
Common use cases
Scenario 1: Verify a new sensor connection The instrumentation technician just installed a pressure sensor on compressor C-04 and configured it in Rela AI as an MQTT source. They open Live Events, filter by the source "pressure-sensor-C04," and watch the feed. Within seconds they start seeing the sensor readings arriving correctly with INFO severity. They confirm the connection is working and close the view.
Scenario 2: Monitoring during plant startup The plant is about to start up after a scheduled 48-hour shutdown. The shift supervisor opens Live Events on a second monitor and leaves it running during the startup. As equipment powers on, they can see in real time whether any alarms are generated, with what severity, and whether the agent is responding correctly. When the main motor generates a temperature WARNING on startup (normal when cold), they can see it and know it is expected.
Scenario 3: Diagnosing repetitive alarms A technician reports that the system is sending too many notifications from sensor TT-12. The maintenance manager opens Live Events, filters by "sensor-TT-12," and watches the feed. Within 2 minutes they see the sensor sending an event every 8 seconds — the throttle is set to 60 seconds but the sensor is generating data very frequently. They identify that the sensor threshold is configured too close to the normal operating value and adjust it.