Real-Time Incident Management Dashboard

This project presents a distributed real-time incident management dashboard designed for collaborative ticket resolution in operational environments. The system enables multiple operators to simultaneously interact with a shared pool of incidents through a web-based dashboard. The architecture follows a three-tier pattern: a Vue.js 3 single-page application serves as the presentation layer, an Express.js API with Socket.IO handles real-time communication and business logic, while a MongoDB Replica Set ensures persistent strongly-consistent state and Redis provides distributed coordination. Key distributed systems patterns implemented include: (i) Redis-based distributed locking and Lua scripts for check-then-act operations, preventing race conditions during ticket assignment; (ii) optimistic locking via document versioning in MongoDB to avoid lost updates during concurrent modifications; (iii) leader election using Redis TTL keys for exclusive execution of background tasks (automatic escalation of stale incidents); (iv) event-driven synchronization of all connected clients without polling, combining MongoDB Change Streams for incident updates with Redis Pub/Sub (via the Socket.IO adapter) for lock and presence events. The entire system is containerized and orchestrated via Kubernetes (validated via Minikube), featuring a 3-node MongoDB StatefulSet for data persistence and failover, Horizontal Pod Autoscaler for the API tier (3–10 replicas), and an Ingress with session affinity.