Coforge's Data Cosmos Pulsar-APS – Agentic Data Production Support

Pulsar-APS (Agentic Production Support) is a Coforge Data Cosmos Artifact that transforms how enterprises manage batch data pipeline operations on Azure. Leveraging a multi-agent AI architecture, Pulsar-APS continuously monitors batch schedules, detects failures within 30 seconds, autonomously diagnoses root causes, and resolves incidents across L1, L1.5, and L2 support tiers — with human-in-the-loop L3 escalation for complex cases. It replaces reactive, human-centric support models with proactive, always-on intelligent agents that learn from every incident and compound institutional knowledge over time. Pulsar-APS is a core component of Coforge Data Cosmos™ — our Innovation Backbone combining platforms, agentic accelerators, and services to enable end-to-end data engineering, BI, governance, and analytics on Azure. The platform deploys three specialized AI agents: • Mr. Glass (L1 Watchman) – Continuous 24/7 monitoring of batch jobs, data pipelines, and infrastructure health. Detects anomalies, triages alerts, creates ITIL-compliant tickets, and escalates based on severity and SLA thresholds. • Mr. QuickFix (L1.5 Fixer) – Pattern-based resolution using a Known Error Database (KEDB). Applies proven playbooks, restarts failed jobs, clears stuck queues, and executes corrective actions autonomously with full audit trails. • Ms. Troubleshoot (L2 Investigator) – Deep diagnostic analysis for complex, multi-system failures. Correlates logs, performs root cause analysis, and generates detailed investigation reports with recommended fixes. • Human-in-the-Loop (L3 Decision Authority) – Steps in only for true judgment calls: infra provisioning, financial approval, and policy sign-off. Pulsar-APS integrates natively with enterprise orchestration tools including Azure Data Factory, Apache Airflow, Control-M, and Autosys. It features built-in ticket management, a knowledge library, real-time operational dashboards, ITIL-compliant reporting, and supports the A2A (Agent-to-Agent) protocol for multi-agent collaboration and MCP (Model Context Protocol) for dynamic skill provisioning. Deployed on Azure using AKS and Azure OpenAI. Quantified Business Outcomes: • 50–60% reduction in 5-year Total Cost of Ownership (TCO) over traditional support models • 40–60% reduction in Mean Time to Resolution (MTTR) • <30 second knowledge search and retrieval latency • 24/7 autonomous monitoring eliminating human fatigue errors The 12-week implementation engagement is delivered in three phases: • Phase 1 – Knowledge Onboarding (Weeks 1–4): Ingest runbooks, SOPs, KEDB, and historical incident data • Phase 2 – Agent Activation (Weeks 5–10): Deploy and configure the three AI agents on Azure, integrate with orchestrators including Azure Data Factory, set up dashboards and alerting • Phase 3 – Supervised Operations & Handover (Weeks 11–12): Monitored autonomous operation with fine-tuning and knowledge transfer to your Azure data teams Target Audience: Data Operations & Support Teams, Platform Engineering Teams, Data Engineering Leads, and IT Service Management Teams adopting or extending Microsoft Azure.