Skip to main content
Platform Feature

Monitoring and
Alerting

Continuous monitoring of every operator, device, and autonomous agent in your environment. The platform evaluates behavioural patterns in real time, detects anomalies through AI-driven baselines, and executes automated responses through configurable triggers.

Behaviour tracking

Real-time alerts and warnings

Every entity is monitored against configurable thresholds. When conditions are breached, alerts are classified by severity and routed through the appropriate notification channels. Critical events trigger immediate escalation.

Real-time pattern evaluation
Behavioural patterns tracked across operators, devices, and agents.
Configurable alert thresholds
Define severity levels and escalation rules per entity type.
Multi-channel notifications
Route alerts to dashboards, email, webhooks, SMS, or integrations.
Live Alert Feed
Monitoring
Warning Forklift FL-09
14:32:07

Operating outside designated zone for 4m 12s

Critical Conveyor C-03
14:28:41

Throughput 62% below hourly baseline

Info Operator J. Ramos
14:15:03

Access pattern deviation detected in Sector 7

Warning Humanoid H-04
14:09:55

Battery level below 15%, charging station occupied

Channels:
DashboardEmailWebhookSMSPagerDutySlack
Trigger Rules
4 active rules
If: Temperature > 85°C on any motor
Armed
Then: Reduce RPM to 60%, notify maintenance lead
If: Humanoid enters restricted zone without clearance
Armed
Then: Halt movement, lock zone perimeter, alert supervisor
If: Conveyor idle > 15 min during active shift
Armed
Then: Log downtime event, reassign pending tasks
If: Air quality index drops below threshold
Testing
Then: Activate ventilation protocol, evacuate zone if critical
Automation

Custom triggers and actions

Define rules that bind specific conditions to automated responses. When a trigger fires, whether from a sensor reading, a geofence breach, or an access anomaly, the platform executes the associated action sequence without manual intervention.

Condition-based trigger definitions
Combine sensor data, entity state, time windows, and logical operators.
Composable action chains
Sequence multiple actions with conditional branching and fallbacks.
Full audit trail
Every trigger activation and action execution logged with timestamps.
Intelligence

Automatic deviance detection

AI models trained on operational baselines continuously evaluate system behaviour and flag deviations. When an entity acts outside its established norms, the platform generates a deviance report with severity classification, probable cause analysis, and recommended remediation.

Unlike rule-based alerting, deviance detection identifies anomalies that no human would think to write a threshold for. The AI learns what normal looks like and flags everything else.

Deviance Reports
AI analysis
Robot Arm RA-02 High
Metric
Cycle time
Baseline
4.2s avg
Current
6.8s avg
Probable: servo motor degradation in joint 3
Warehouse Zone B Medium
Metric
Pick rate
Baseline
142/hr
Current
89/hr
Probable: routing congestion from reslotting
HVAC Unit AC-11 Low
Metric
Energy consumption
Baseline
12.4 kWh
Current
19.1 kWh
Probable: filter blockage reducing airflow efficiency
Capabilities

What Miradoris enables

Real-time behaviour tracking

Every operator, humanoid, device, and environmental sensor is monitored against established behavioural baselines in real time.

Configurable severity levels

Define thresholds for Info, Warning, and Critical alerts per entity type. Escalation rules route alerts to the appropriate response channel.

Geofence and zone monitoring

Track entity positions relative to defined zones. Receive alerts when entities enter restricted areas or leave designated operating zones.

Baseline learning

AI builds behavioural profiles for each entity, process, and environment over time. Profiles adapt as operational patterns evolve.

Anomaly scoring

Every deviation is scored by severity with confidence intervals. High-confidence anomalies trigger automatic responses, low-confidence events are queued for review.

Root cause correlation

AI correlates deviations with system state, environmental conditions, and recent changes to identify probable root causes automatically.

Comparison

Approach analysis

Manual Monitoring

Strengths
No technology investment
Human judgment for context
Simple to start
Limitations
Cannot scale to large environments
Fatigue reduces detection accuracy
No real-time pattern analysis
No audit trail

Traditional SCADA Alerts

Strengths
Proven threshold-based alerting
Fast response for known conditions
Industry standard protocols
Limitations
Rule-based only, no AI detection
No behavioural baselines
Cannot detect novel anomalies
Siloed from operational context

Miradoris

Recommended
Strengths
AI-driven deviance detection
Behavioural baseline learning
Unified monitoring across all entity types
Custom triggers with composable actions
Full audit trail with root cause analysis
FAQ

Frequently asked questions

What types of entities can be monitored?

Any entity modeled in the Miradoris ontology can be monitored: humanoids, AGVs, conveyors, sensors, operators, environmental systems, and custom device types. Each entity type has configurable behavioural baselines and alert thresholds.

How does AI deviance detection differ from threshold alerts?

Threshold alerts fire when a single value crosses a predefined limit. AI deviance detection learns normal behavioural patterns across multiple dimensions and flags deviations that may not breach any individual threshold but represent abnormal composite behaviour.

Can triggers execute actions on physical systems?

Yes. Trigger actions can send commands to PLCs via OPC UA, dispatch humanoid instructions via ROS 2 or REST, adjust device parameters via MQTT, and activate safety protocols. Actions are logged with full audit context.

How are notification channels configured?

Channels are configured per alert severity and entity type. A critical alert on a safety system might route to SMS and PagerDuty, while an informational pattern deviation routes to the dashboard only. Routing rules are managed through the platform interface.

How quickly are anomalies detected?

Behavioural evaluations run continuously against streaming data. Detection latency depends on the metric type: real-time sensor anomalies are flagged within seconds, while trend-based deviations require sufficient observation windows for statistical confidence.

Can monitoring rules be tested before deployment?

Yes. Triggers can be set to a Testing state where they evaluate conditions and log results without executing actions. This allows validation against live data before arming the rule for production response.

Be among the first

We are looking for partners willing to test Miradoris in real operational environments. Early adopters get priority access to the platform at significantly reduced rates.

We'll review your request and follow up. No unsolicited contact.