Event-driven architecture (EDA) has emerged as a pivotal paradigm for real-time data processing in distributed systems. As modern applications demand low-latency responses, scalability, and fault tolerance, event-driven systems enable asynchronous communication, improving responsiveness and system efficiency. This paper explores the core principles of event-driven architectures, including event sourcing, choreography, and orchestration, and examines their integration with microservices, distributed databases, and cloud-native technologies. It discusses key challenges such as event ordering, idempotency, fault tolerance, and scalability in large-scale distributed systems. Additionally, it presents industry use cases demonstrating effective implementations of EDA for streaming analytics, financial transactions, and IoT data processing. A comparative analysis of event brokers such as Apache Kafka, RabbitMQ, and AWS EventBridge highlights their trade-offs in terms of performance, reliability, and scalability. The paper concludes with best practices for designing and optimizing event-driven systems, offering insights into architectural patterns that enhance resiliency and maintainability in real-time data pipelines.