Skip to main content

ML Engine API

Complete API reference for predictions, batch processing, and model management.

Base URL​

https://[your-app].azurecontainerapps.io

Replace [your-app] with your Azure Container App name.

Authentication​

Pay-As-You-Go: No authentication (runs in your private network)
BYOL: License key via environment variable

Network Security:

  • Deploy in your VNet
  • Configure ingress rules
  • Use API Gateway for external access

Endpoints​

Health Check​

GET /v1/health

Response:

{
  "status": "healthy",
  "version": "1.0.0",
  "models_loaded": true,
  "uptime_seconds": 3847.2
}

Single Prediction​

POST /v1/predict
Content-Type: application/json

Request:

{
  "vehicleId": "TRUCK-001",
  "timestamp": "2026-01-12T10:00:00Z",
  "coolant": 95.5,
  "engineOilTemp": 102.3,
  "vehicleSpeed": 65.0,
  "engine_speed": 2100,
  "battery_voltage": 13.8,
  "fuel_level": 75.0,
  "obdDistance": 45000,
  "mil_lamp": "LAMP_OFF"
}

Response:

{
  "vehicleId": "TRUCK-001",
  "timestamp": "2026-01-12T10:00:00Z",
  "processingTimeMs": 42.5,
  "predictions": {
    "healthScore": 87.5,
    "maintenanceNeeded": false,
    "maintenanceProbability": 0.12,
    "remainingUsefulLife": {
      "engine": 180,
      "brakes": 120,
      "battery": 240
    },
    "anomalyScore": 0.08,
    "anomalyDetected": false
  },
  "alerts": [],
  "insights": [
    "Vehicle performing normally",
    "Oil change due in 45 days"
  ]
}

Batch Prediction​

POST /v1/predict/batch
Content-Type: application/json

Request:

{
  "vehicles": [
    {"vehicleId": "V001", "coolant": 95.5, ...},
    {"vehicleId": "V002", "coolant": 98.0, ...}
  ]
}

Limit: 1,000 vehicles per request

Response:

{
  "count": 2,
  "processingTimeMs": 150.5,
  "predictions": [
    {...},
    {...}
  ]
}

List Models​

GET /v1/models

Response:

{
  "models": [
    {
      "name": "health_predictor",
      "type": "regressor",
      "accuracy": 0.94
    },
    {
      "name": "rul_lstm",
      "type": "deep_learning",
      "mae_days": 5.2
    }
  ]
}

Input Schema​

GET /v1/schema/input

Returns JSON schema for input validation.

Output Schema​

GET /v1/schema/output

Returns JSON schema for response format.

Schema Aliases​

GET /v1/schema/aliases

Returns all supported field aliases for auto-mapping:

{
  "description": "Field aliases that are automatically mapped",
  "aliases": {
    "engine_speed": ["rpm", "engine_rpm", "engineRPM"],
    "coolant": ["coolant_temp", "engine_temp", "ect"],
    "battery_voltage": ["battery", "bat_v", "voltage"]
  }
}

Schema Statistics​

GET /v1/schema/statistics

Returns processing statistics for monitoring data quality:

{
  "total_processed": 1500,
  "fully_standard": 1200,
  "required_mapping": 300,
  "standard_compliance_rate": 80.0
}

API Documentation​

GET /docs

Interactive Swagger/OpenAPI documentation.

Performance​

MetricExpected Value
Health check< 50ms
Single prediction< 500ms
Batch (100 vehicles)< 5 seconds
Concurrent requests1,000/sec per replica

Error Handling​

All errors return JSON:

{
  "error": {
    "code": "invalid_input",
    "message": "Missing required field: vehicleId"
  }
}

HTTP Status Codes:

  • 200 - Success
  • 400 - Bad request
  • 500 - Server error
  • 503 - Service unavailable

Rate Limiting​

Pay-As-You-Go: No limits (you control scaling)
BYOL: Unlimited

Your Azure Container Apps handles auto-scaling.

Next Steps​