Alert Rules

A dashboard only helps when someone is looking at it. Alert rules watch your data day and night and tell you when something needs attention: a temperature too high, a machine that stopped, a sensor that went silent.

An alert rule checks a data query on an interval. When the condition is met, an alert becomes active and a message can be sent through an Alert Output: email, Telegram or a webhook to another system. When the condition recovers, the alert is resolved.

Create an alert rule

Go to the Alert Rules page and click:

Name and description

Give the rule a clear name — it is used in the alert messages. For example: Oven temperature too high beats alert_rule_7.

Data

Build the data query the same way as a dashboard panel: pick the tag, the aggregation and any calculation steps. A live preview shows the current value, so you see immediately what the rule will evaluate.

Condition

Choose when the alert should fire:

Threshold — the value crosses a limit. Pick an operator (>, <, >=, <=, ==, !=) and a value. Example: temperature > 80.
Range — the value leaves (or enters) a band between a minimum and a maximum. Example: line speed outside 90–110% of the setpoint.
No data — no datapoints arrived within the lookback window. The condition for “is my sensor still alive?”. Put a No data rule on your most important tags — a silent sensor is easy to miss.

Timing

Evaluation Interval — how often the rule is checked, in seconds.
Lookback — how far the rule looks back for data, in seconds. Make the lookback comfortably larger than the deduplication timeout of the tag, so heartbeat datapoints are always inside the window.

Output

Select an Alert Output, or create one on the spot. Without an output the alert is only visible in UpBlue.

Optionally override the message. Wildcards are filled in when the alert fires: {{value}}, {{threshold}}, {{operator}}, {{rule_name}}, {{rule_description}}, {{timestamp}}

For example: Oven 2 is at {{value}} °C (limit {{threshold}}). Check the burner.

A message that says what to do beats a message that only says what happened.

Testing

On the rule page, click: to evaluate the rule immediately and see the result.

The Evaluation Log shows every check the rule did and its outcome. When a rule does not fire as expected, this log tells you why.

Alert events

Every rule keeps its history: which alerts are Active right now, which are Resolved and how long they lasted. On a dashboard, the Alert list panel shows this where everyone can see it.

💡

An alert rule can also record its alert periods as operations, with a duration per alert. For tracking production runs this is not the way — use rule events instead. Use it when you want alert history as analyzable data.

Active switch

A rule can be switched off without deleting it, for example during a planned maintenance stop. Inactive rules are clearly marked in the list.