Clinical analytics platforms continuously ingest streaming data from electronic health systems, medical devices, laboratory results, and administrative sources. As hospitals modernize their infrastructures using containerized micro services and real-time ETL, incomplete visibility into data flows leads to silent failures, undetected schema drift, compliance risks, and unreliable downstream models. Observability-centric DevOps offers a structured way to track data lineage, monitor transformation integrity, and validate runtime service behavior. This work presents an architecture integrating FHIR-based ingestion, Kafka streaming, microservices ETL, and a unified observability layer incorporating logging, distributed tracing, metrics, and runtime assurance for healthcare pipelines. The proposed framework elevates reliability by enabling proactive debugging, semantic data integrity guarantees, and automated error localization across heterogeneous clinical systems.