Table 8 Scalability and Resource Metrics Under Load Replay.
From: Hybrid MLOps framework for automated lifecycle management of adaptive phishing detection models
Metric | Result |
|---|---|
Load replay volume | 5 million labeled URLs |
Ingestion rate | \(\approx\) 2,300 requests per second |
p99 inference latency | 41.6 milliseconds |
Horizontal Pod Autoscaler (HPA) scaling behavior | Scaled from 3 to 12 pods (CPU target = 70%) |
GPU utilization (peak during retraining) | 74% |
Storage overhead after 60 days | \(\le\) 1.4 TB (MinIO lifecycle-managed) |
