Logging & Monitoring

Role in the Project

Provides observability for system health, API performance, and AI processing efficiency.

Strengths & Weaknesses

Strengths:

• Real-time metrics collection for AI and backend performance
• Configurable dashboards for anomaly detection

Weaknesses:

• Requires fine-tuning of alerts to avoid false positives
• Large-scale logging may introduce storage overhead

Available Technologies & Comparison

• Logging: ELK Stack (Elasticsearch, Logstash, Kibana) vs. Loki (lightweight alternative)
• Monitoring: Prometheus (real-time), Grafana (visualization)

Chosen Approach

• Prometheus for metrics collection (low-latency queries)
• Grafana for custom dashboards (integrates well with Prometheus and Kafka event logs)

Example of Prometheus metric for AI inference time:

http_server_requests_seconds_count{endpoint="/predict"}

This architecture ensures EYNTRY's backend remains secure, scalable, and efficient for AI-driven vision applications.