Skip to main content
The Operator Input Log panel allows maintenance personnel to record real-world maintenance events that serve as ground-truth labels for supervised machine learning.

​
Purpose

Operator logs create a bridge between sensor anomalies and actual maintenance activities, enabling:
  • Supervised ML Training: Correlate sensor patterns with known failure modes
  • Audit Trail: Document maintenance history for compliance and analysis
  • Chart Overlay: Visualize maintenance events on the sensor telemetry chart
  • Failure Mode Analysis: Identify which sensor patterns precede specific failure types
All operator logs are persisted to InfluxDB in the maintenance_logs measurement within the sensor_data bucket.

​
Logging a Maintenance Event

The form includes the following fields:
1

Asset ID

Field: Read-only (pre-filled with Motor-01)Purpose: Identifies which asset the maintenance was performed on.
2

Event Type

Field: Dropdown (required)Options:
  • Preventive Lubrication — Scheduled lubrication activity
  • Preventive Cleaning — Scheduled cleaning activity
  • Preventive Inspection — Routine inspection
  • Corrective Bearing Replacement — Bearing replacement due to wear
  • Corrective Alignment — Motor/pump alignment adjustment
  • Corrective Electrical — Electrical system repair
  • Status Calibration — System recalibration event
  • Status Restart — System restart or reboot
3

Severity

Field: Dropdown (required)Options:
  • LOW — Minor maintenance, no production impact
  • MEDIUM — Moderate issue, scheduled downtime
  • HIGH — Urgent repair, unplanned downtime
  • CRITICAL — Emergency failure, immediate action
4

Description

Field: Text area (required, max 1000 characters)Purpose: Free-text description of the maintenance activity.Example:
Replaced main bearing due to excessive vibration.
Part #SKF-6205-2RS. Vibration reduced from 0.35g to 0.14g post-replacement.
5

Event Date & Time

Field: Datetime picker (optional)Purpose: Allows backdating logs for historical events. If left empty, the current timestamp is used.Use Case: Log a bearing replacement that occurred 2 days ago to correlate with past sensor spikes.

​
Example Log Entry

Bearing Replacement

Asset: Motor-01
Type: Corrective Bearing Replacement
Severity: HIGH
Description: Replaced main bearing due to excessive vibration. Part #SKF-6205-2RS.
Timestamp: 2026-01-31T10:30:00Z

​
How Logs Are Stored

When you submit a log entry, it is written to InfluxDB with the following schema:
Name: maintenance_logsBucket: sensor_data (same as sensor telemetry)

​
InfluxDB Point Example

maintenance_logs,asset_id=Motor-01,event_type=CORRECTIVE_BEARING_REPLACEMENT,severity=HIGH 
description="Replaced main bearing due to excessive vibration. Part #SKF-6205-2RS.",
event_id="550e8400-e29b-41d4-a716-446655440000" 
1738322400000000000

​
Chart Overlay (Future Enhancement)

Operator logs can be queried and displayed as vertical markers on the sensor telemetry chart: 
// Fetch logs for the past 24 hours
GET /api/logs?hours=24&asset_id=Motor-01

// Response:
{
  "logs": [
    {
      "timestamp": "2026-01-31T10:30:00Z",
      "event_type": "CORRECTIVE_BEARING_REPLACEMENT",
      "severity": "HIGH",
      "description": "Replaced main bearing..."
    }
  ],
  "count": 1
}
These logs can be overlaid on the chart as orange diamonds at the corresponding timestamps, allowing operators to see which maintenance events align with sensor anomalies.

​
ML Training Use Case

Operator logs serve as supervised labels for training failure prediction models:

​
Example: Bearing Failure Prediction

  1. Collect Data: 6 months of sensor data + operator logs
  2. Label Events: Extract all CORRECTIVE_BEARING_REPLACEMENT events
  3. Feature Extraction: Pull sensor data from 24 hours before each replacement
  4. Train Classifier: Learn patterns that precede bearing failures
  5. Deploy Model: Predict bearing failures 7-14 days in advance
The /api/logs endpoint allows querying logs by time range and asset, making it easy to export data for offline ML training.

​
API Reference

​
Create Log Entry

POST /api/log
Content-Type: application/json

{
  "asset_id": "Motor-01",
  "event_type": "CORRECTIVE_BEARING_REPLACEMENT",
  "severity": "HIGH",
  "description": "Replaced main bearing due to excessive vibration. Part #SKF-6205-2RS.",
  "timestamp": "2026-01-31T10:30:00Z"  // Optional
}
Response (201 Created): 
{
  "event_id": "550e8400-e29b-41d4-a716-446655440000",
  "asset_id": "Motor-01",
  "event_type": "CORRECTIVE_BEARING_REPLACEMENT",
  "severity": "HIGH",
  "timestamp": "2026-01-31T10:30:00Z",
  "message": "Log entry created successfully for Motor-01"
}

​
Retrieve Logs

GET /api/logs?hours=24&limit=50&asset_id=Motor-01
Response: 
{
  "logs": [
    {
      "timestamp": "2026-01-31T10:30:00Z",
      "asset_id": "Motor-01",
      "event_type": "CORRECTIVE_BEARING_REPLACEMENT",
      "severity": "HIGH",
      "description": "Replaced main bearing due to excessive vibration. Part #SKF-6205-2RS.",
      "event_id": "550e8400-e29b-41d4-a716-446655440000"
    }
  ],
  "count": 1,
  "query_params": {
    "hours": 24,
    "limit": 50,
    "asset_id": "Motor-01"
  }
}

​
List Event Types

GET /api/log/types
Response: 
{
  "event_types": [
    {
      "value": "PREVENTIVE_LUBRICATION",
      "category": "PREVENTIVE",
      "label": "Preventive Lubrication"
    },
    {
      "value": "CORRECTIVE_BEARING_REPLACEMENT",
      "category": "CORRECTIVE",
      "label": "Corrective Bearing Replacement"
    }
    // ... more event types
  ],
  "severities": ["LOW", "MEDIUM", "HIGH", "CRITICAL"]
}

​
InfluxDB Persistence Benefits

Time-Series Alignment

Logs share the same timestamp domain as sensor data, enabling precise correlation

Indexed Tags

Fast queries by asset_id, event_type, and severity using InfluxDB’s tag indexing

Retention Policies

Configure automatic data expiration for older logs (e.g., keep 2 years)

Flux Query Integration

Use Flux to join sensor data with maintenance logs for correlation analysis

​
Best Practices

1

Log Immediately

Record maintenance events as they happen to ensure timestamp accuracy.
2

Be Specific

Include part numbers, measurements (e.g., “vibration reduced from 0.35g to 0.14g”), and technician notes.
3

Use Correct Severity

Reserve CRITICAL for emergency failures. Use MEDIUM for routine repairs.
4

Backdate When Needed

Use the datetime picker to log historical events that weren’t recorded in real-time.

​
Next Steps

Dashboard Overview

Learn about the overall dashboard layout and features

Metrics & Health Scoring

Understand how sensor metrics are displayed and calculated

Build docs developers (and LLMs) love

Get started for freeTalk to us