January 21, 2024 5 min read

Data Activator (Reflex) Alerts: Event-Driven Actions in Fabric

Data Activator Reflex Microsoft Fabric Alerts Automation

Data Activator, codenamed Reflex, brings event-driven automation to Microsoft Fabric. It monitors data in real-time and triggers actions when conditions are met - no code required for basic scenarios.

Core Concepts

Data Activator consists of:

Objects: Represent real-world entities (devices, customers, orders)
Triggers: Define conditions that activate actions
Actions: What happens when triggers fire (email, Teams, Power Automate)

Setting Up Data Activator

Creating Your First Reflex

# Conceptual API for Reflex configuration
reflex_config = {
    "name": "Production Monitoring Reflex",
    "description": "Monitor production line sensors",
    "data_sources": [
        {
            "type": "KqlDatabase",
            "database": "TelemetryDB",
            "table": "SensorReadings"
        },
        {
            "type": "Eventstream",
            "eventstream": "production-events"
        }
    ]
}

Defining Objects

Objects represent the things you want to monitor:

# Object definition
object_definition = {
    "name": "ProductionLine",
    "id_column": "line_id",
    "properties": [
        {
            "name": "CurrentTemperature",
            "source_column": "temperature",
            "aggregation": "latest"
        },
        {
            "name": "AverageTemperature",
            "source_column": "temperature",
            "aggregation": "avg",
            "window": "5m"
        },
        {
            "name": "AlertCount",
            "source_column": "alert_flag",
            "aggregation": "sum",
            "window": "1h"
        },
        {
            "name": "Status",
            "source_column": "status",
            "aggregation": "latest"
        }
    ]
}

Trigger Patterns

Pattern 1: Threshold Alert

trigger:
  name: High Temperature Alert
  object: ProductionLine
  condition:
    property: CurrentTemperature
    operator: greaterThan
    value: 85
  action:
    type: email
    recipients:
      - operations@company.com
    subject: "High Temperature Alert: {{object.line_id}}"
    body: |
      Production line {{object.line_id}} has exceeded temperature threshold.

      Current Temperature: {{object.CurrentTemperature}}°C
      Threshold: 85°C
      Time: {{trigger.timestamp}}

      Please investigate immediately.

Pattern 2: Trend-Based Alert

trigger:
  name: Temperature Rising Alert
  object: ProductionLine
  condition:
    type: trend
    property: AverageTemperature
    direction: increasing
    threshold: 10  # 10% increase
    window: 30m
  action:
    type: teams
    channel: production-alerts
    message: |
      **Temperature Trend Alert**
      Line: {{object.line_id}}
      Trend: Rising {{trend.percent_change}}% over 30 minutes
      Current: {{object.CurrentTemperature}}°C

Pattern 3: Absence Detection

trigger:
  name: Device Offline Alert
  object: Sensor
  condition:
    type: absence
    property: LastReading
    duration: 5m
  action:
    type: email
    recipients:
      - iot-support@company.com
    subject: "Sensor Offline: {{object.sensor_id}}"
    body: |
      Sensor {{object.sensor_id}} has not reported data for 5 minutes.

      Last reading: {{object.LastReading}}
      Last seen: {{object.LastSeenTimestamp}}

Pattern 4: Compound Conditions

trigger:
  name: Critical Production Alert
  object: ProductionLine
  condition:
    type: compound
    operator: and
    conditions:
      - property: CurrentTemperature
        operator: greaterThan
        value: 80
      - property: Status
        operator: equals
        value: "Running"
      - property: AlertCount
        operator: greaterThan
        value: 5
  action:
    type: powerAutomate
    flow: "Critical Production Response"
    parameters:
      line_id: "{{object.line_id}}"
      temperature: "{{object.CurrentTemperature}}"
      alert_count: "{{object.AlertCount}}"

Pattern 5: State Change Detection

trigger:
  name: Status Change Alert
  object: ProductionLine
  condition:
    type: stateChange
    property: Status
    from: "Running"
    to: "Stopped"
  action:
    type: teams
    channel: production-ops
    message: |
      **Production Line Status Change**
      Line: {{object.line_id}}
      Previous Status: Running
      New Status: Stopped
      Time: {{trigger.timestamp}}

      Investigate cause of stoppage.

Action Types

Email Actions

email_action = {
    "type": "email",
    "recipients": ["team@company.com"],
    "cc": ["manager@company.com"],
    "subject": "Alert: {{trigger.name}}",
    "body": """
    Alert Details:
    - Object: {{object.id}}
    - Condition: {{trigger.condition}}
    - Value: {{trigger.value}}
    - Time: {{trigger.timestamp}}
    """,
    "priority": "high"
}

Teams Actions

teams_action = {
    "type": "teams",
    "webhook_url": "https://company.webhook.office.com/...",
    "adaptive_card": {
        "type": "AdaptiveCard",
        "body": [
            {
                "type": "TextBlock",
                "text": "Alert: {{trigger.name}}",
                "weight": "bolder",
                "size": "large"
            },
            {
                "type": "FactSet",
                "facts": [
                    {"title": "Object", "value": "{{object.id}}"},
                    {"title": "Value", "value": "{{trigger.value}}"},
                    {"title": "Time", "value": "{{trigger.timestamp}}"}
                ]
            }
        ],
        "actions": [
            {
                "type": "Action.OpenUrl",
                "title": "View Dashboard",
                "url": "https://app.fabric.microsoft.com/..."
            }
        ]
    }
}

Power Automate Integration

power_automate_action = {
    "type": "powerAutomate",
    "flow_name": "Handle Production Alert",
    "trigger_url": "https://prod-xx.westus.logic.azure.com/...",
    "parameters": {
        "alert_type": "{{trigger.name}}",
        "object_id": "{{object.id}}",
        "current_value": "{{trigger.value}}",
        "threshold": "{{trigger.threshold}}",
        "timestamp": "{{trigger.timestamp}}"
    }
}

Real-World Example: E-Commerce Monitoring

# Complete Reflex configuration for e-commerce monitoring

ecommerce_reflex = {
    "name": "E-Commerce Monitoring",

    "objects": [
        {
            "name": "Order",
            "id_column": "order_id",
            "properties": [
                {"name": "Status", "column": "status", "agg": "latest"},
                {"name": "TotalValue", "column": "total", "agg": "sum"},
                {"name": "ItemCount", "column": "items", "agg": "count"}
            ]
        },
        {
            "name": "Product",
            "id_column": "product_id",
            "properties": [
                {"name": "StockLevel", "column": "stock", "agg": "latest"},
                {"name": "SalesVelocity", "column": "sales", "agg": "sum", "window": "1h"}
            ]
        },
        {
            "name": "Customer",
            "id_column": "customer_id",
            "properties": [
                {"name": "OrderCount", "column": "order_id", "agg": "count", "window": "24h"},
                {"name": "TotalSpend", "column": "total", "agg": "sum", "window": "24h"}
            ]
        }
    ],

    "triggers": [
        {
            "name": "High Value Order",
            "object": "Order",
            "condition": {"property": "TotalValue", "op": ">", "value": 1000},
            "action": {"type": "email", "to": "vip-sales@company.com"}
        },
        {
            "name": "Low Stock Alert",
            "object": "Product",
            "condition": {
                "type": "compound",
                "op": "and",
                "conditions": [
                    {"property": "StockLevel", "op": "<", "value": 10},
                    {"property": "SalesVelocity", "op": ">", "value": 5}
                ]
            },
            "action": {"type": "teams", "channel": "inventory"}
        },
        {
            "name": "VIP Customer Activity",
            "object": "Customer",
            "condition": {"property": "TotalSpend", "op": ">", "value": 5000},
            "action": {"type": "powerAutomate", "flow": "VIP Customer Handler"}
        }
    ]
}

Best Practices

1. Alert Fatigue Prevention

# Configure alert throttling
trigger_config = {
    "name": "Temperature Alert",
    "condition": {"property": "Temperature", "op": ">", "value": 80},
    "throttling": {
        "max_alerts_per_hour": 2,
        "cooldown_minutes": 30,
        "group_by": "object_id"
    }
}

2. Escalation Patterns

# Multi-level escalation
escalation_pattern = {
    "levels": [
        {
            "delay": "0m",
            "action": {"type": "email", "to": "on-call@company.com"}
        },
        {
            "delay": "15m",
            "condition": "not_acknowledged",
            "action": {"type": "teams", "channel": "urgent"}
        },
        {
            "delay": "30m",
            "condition": "not_acknowledged",
            "action": {"type": "phone", "to": "manager"}
        }
    ]
}

3. Testing Triggers

# Test trigger logic before deployment
def test_trigger(trigger_config: dict, test_data: list[dict]) -> list[dict]:
    """Test trigger conditions against sample data."""

    results = []

    for record in test_data:
        triggered = evaluate_condition(trigger_config["condition"], record)

        results.append({
            "record": record,
            "triggered": triggered,
            "trigger_name": trigger_config["name"]
        })

    return results

# Test data
test_data = [
    {"line_id": "L1", "temperature": 75, "status": "Running"},
    {"line_id": "L2", "temperature": 90, "status": "Running"},
    {"line_id": "L3", "temperature": 85, "status": "Stopped"}
]

results = test_trigger(high_temp_trigger, test_data)

Conclusion

Data Activator brings proactive intelligence to your data platform:

Define objects that represent business entities
Create triggers for conditions that matter
Configure actions to respond automatically
Prevent alert fatigue with throttling and escalation

The combination of Eventstreams, KQL databases, and Data Activator creates a complete real-time intelligence stack - from data ingestion to automated response.