Monitoring
Inference metrics, model performance, and system health
Total Predictions
-
Avg Latency
-
Fraud Rate
-
Accuracy
N/A
Inference Service Health
Real-time status of model inference service
Service degraded
No model status available
Production Models
Currently deployed model versions
No production models deployed
Drift Detection
Automated monitoring for data and concept drift
Metrics will appear here once the inference service starts sending prediction data.
Configure METRICS_PUSH_URL on ml-inference to enable metrics collection.