producer

Generate synthetic transaction streams for Kafka-based fraud detection by producing configurable phase-based distributions with per-user behavioral state to the raw-transactions topic. The producer applies deterministic event generation and anomaly label rules for model evaluation.

Drift simulation in anomaly detection pipelines works by cycling through controlled normal, drift, and attack phases with phase-transition signaling. This allows you to test model robustness against shifting transaction distributions and apply ground-truth anomaly labels.

Publish synthetic Kafka events with retry-safe topic initialization by managing startup topic creation with built-in retry handling. The producer ensures reliable event generation to raw-transactions while tracking throughput via Prometheus metrics and health endpoints.

Apply rule-based anomaly labels to synthetic financial transaction data to support drift detection and model evaluation. The producer generates ground-truth labels alongside controlled normal, drift, and attack behaviors for streaming anomaly detection systems.

Monitor synthetic transaction throughput in a fraud detection pipeline using integrated Prometheus metrics and health endpoints. The producer provides operational visibility to track event generation rates and behavior phases across the streaming pipeline.

Kafka producers fail to initialize topics for anomaly detection streams when topic creation lacks retry handling. The producer service manages startup topic creation with retry-safe publishing to ensure raw-transactions and phase-transition topics are established reliably.