Support for horizontal scaling of OpenMRS instances.

OpenMRS has historically operated as a single-instance application, creating a "hard ceiling" for scaling and a single point of failure. As global healthcare networks grow, there is an urgent need for High Availability (HA). While foundational work on distributed caching and storage was introduced in version 2.8.x, several critical gaps remain: lack of distributed session management, race conditions during cluster startup, duplicate execution of scheduled tasks, and inconsistent search indexes across nodes. This project aims to complete the "Cluster-Ready" roadmap for OpenMRS 3.x. Drawing inspiration from proven architectures like Confluence Data Center, I will implement a Zero-Footprint, Multi-Replica framework. I hope to do this by Implementing a formal "Local Home vs. Shared Home" split to manage node-specific vs. cluster-wide data. I want to make use of distributed State Introducing Redis and MySQL-backed session management so users remain logged in even if routed to different replicas. Leveraging Infinispan to implement distributed task locking (preventing duplicate background jobs) and a "Cluster Safety" mechanism to detect and mitigate split-brain scenarios. Transitioning Hibernate Search to a shared OpenSearch/Elasticsearch backend to ensure search consistency across the entire cluster. Project Deliverables: A safe startup sequence for containerized environments (Kubernetes/Helm). Configuration-driven support for Redis (Redisson) and JDBC session persistence. A wrapper for the OpenMRS SchedulerService to ensure tasks run on exactly one node. Audit and migration of O3 modules (FHIR2, Webservices.rest) to the distributed cache manager. Application-level wiring for shared full-text search indexes. A robust integration testing framework using Testcontainers to validate cluster behavior in CI/CD.