Skip to main content

6 Events

An event in TDengine IDMP is a discrete operational occurrence with a defined start time, end time, and duration — the digital record that something happened. A pump tripped, a temperature exceeded its limit, a batch phase completed, a maintenance window began. This concept is equivalent to event frames in OSIsoft PI System, one of the most powerful ideas in industrial data management.

Raw sensor streams tell you what a value was at a given moment; events tell you what was happening operationally — and for how long. Instead of searching through millions of data points to find when a compressor ran in surge, you query the structured event record that already captured it.

Events in the AI Era​

Events matter even more as AI becomes central to industrial operations. AI and machine learning systems work best when data is structured and contextualized — and that is exactly what events provide. Instead of feeding a model millions of raw sensor readings, you give it structured records: "Compressor Surge, Start: 10:23:15, Duration: 12 seconds, Severity: High." That context is what turns signal data into something a model can reason about.

Events directly enable the industrial AI use cases that matter most: training predictive maintenance models, powering anomaly detection, performing root cause analysis, and driving AI agents that can reason about what is happening in a plant. Every one of these requires knowing not just what values were measured, but what operational conditions those values represented — and for how long. Events are the bridge between continuous time-series data and the operational intelligence that AI systems need to be useful.

Event Lifecycle​

Events in TDengine IDMP are always generated automatically by analysis rules associated with an element. The full lifecycle is:

Event Template (defined in Libraries)
        ↓
Analysis (configured on an element, references the template)
        ↓
Event (generated automatically when the analysis condition is met)
        ↓
Notification (optionally sent to configured contact points)

Every event must be based on an event template, which defines the naming pattern, severity level, categories, custom attribute schema, and acknowledgment requirements. Event templates are managed under Libraries → Event Template.

Standard Event Fields​

Every event carries the following standard fields:

FieldDescription
NameDisplay name generated from the naming pattern in the event template
Start TimeWhen the event began
End TimeWhen the event ended (blank if still active)
DurationElapsed time between start and end
TemplateThe event template this event was created from
Severity LevelSeverity category (Critical, Major, Minor, Warning, Normal)
Reason CodeOptional code identifying the cause
CategoriesTags for filtering and grouping
DescriptionFree-text description
Associated ElementThe element that generated this event
Associated AnalysisThe analysis rule that triggered this event
StatusWhether the event has been acknowledged

In addition to these standard fields, an event can carry custom attributes — named values recorded at the time of the event, such as the peak temperature during an exceedance or the batch ID at the time of a fault. Custom attributes are defined in the event template.

What's Covered in This Chapter​