From SAPinsider Las Vegas 2025: How Keurig Dr. Pepper Streamlined SAP Monitoring with Splunk and PowerConnect

When Keurig Dr Pepper merged its operations, it faced the significant challenge of unifying two separate SAP environments with different infrastructure, monitoring tools, and support processes. With five ERP systems, multiple operating systems, and databases spread across cloud and on-premise environments, the company needed a centralized approach to SAP monitoring.

At SAPinsider Las Vegas 2025, Jason Nasse, SAP Operations Manager at Keurig Dr. Pepper, described the company’s system monitoring transformation from a manual, siloed monitoring process to a real-time, AI-driven system using Splunk and PowerConnect, which has delivered significant operational efficiencies, reducing downtime, improving root cause analysis (RCA), and cutting alert fatigue.

Following the merger, Keurig Dr Pepper inherited two distinct SAP environments:

• Keurig’s SAP system (live since 2017) had 15 applications across 135 servers, running on Oracle, Hana, SQL Server, and Sybase databases with a mix of Windows and Linux.
• Dr Pepper’s SAP landscape, in place for 25+ years, had 20 different applications with AIX, Red Hat Linux, and Windows servers, supporting the same database diversity.

The lack of integration between these systems resulted in four separate managed service providers supporting different data centers, job schedulers, and monitoring tools. The SAP teams relied on manual daily checks, consuming four hours of work from two employees every day—a highly inefficient and reactive process.

Without a unified monitoring solution, issues were often missed, leading to frequent critical incidents and ineffective root cause analysis (RCA).  To address these inefficiencies, Keurig Dr Pepper’s leadership set two key goals: consolidate managed service providers and migrate all SAP systems to the Google Cloud Platform (GCP) within a year; and automate SAP monitoring and create a single, real-time visibility layer across all SAP landscapes.

By March 2020, despite the onset of COVID-19, Keurig Dr. Pepper successfully completed its cloud migration. The company then focused on streamlining SAP monitoring, initially attempting SAP Solution Manager before transitioning to Splunk with PowerConnect to gain real-time analytics and AI-driven alerts. Key enhancements included:

• Unified SAP Monitoring – Integrated SAP logs, OS alerts, and network monitoring into one central platform.
• Near Real-Time Data Visibility – Enabled SAP teams to track performance issues live, improving RCA accuracy.
• Automated Incident Management – Directly linked to ServiceNow for automatic ticket generation and prioritization.
• Custom Dashboards and KPIs – Allowed teams to create tailored visual dashboards for SAP application health.
• Alert Fatigue Reduction – Used AI-driven threshold adjustments to eliminate false positives and prioritize critical alerts.
• Faster RCA Turnaround – Reduced RCA investigation times by 70%, significantly improving problem resolution.
• Proactive Issue Resolution – Shifted from a reactive to a predictive maintenance model, reducing SAP downtime incidents.

Nasse summarized the key results and business impacts as follows:

Operational Efficiency Gains

• Mean Time to Resolution (MTTR) reduced by 70%, allowing the SAP team to resolve issues much faster.
• SAP incident priority levels dropped by 30%, meaning fewer high-priority disruptions.

Significant Downtime Reduction

• SAP maintenance window incidents dropped to near zero—previous failures in monthly updates caused production losses, particularly in Keurig’s pod manufacturing.
• Pod production losses due to SAP downtime completely eliminated in 2023.

Reduced Alert Fatigue and Improved RCA

• SAP support teams cut false alerts from 800 per month to under 300, reducing wasted effort on non-issues.
• 10-day SLA for RCA resolution was significantly improved, enabling teams to diagnose root causes in near real-time.

Integration with Business Processes

• Integrated SAP iDocs monitoring with business operations, allowing customer service teams to track order failures and resolve issues before they impacted customers.
• Expanded SAP monitoring to other enterprise applications, improving data flow tracking across MuleSoft, DSD, and other critical platforms.

Keurig Dr. Pepper is now focused on enhancing predictive analytics and DevSecOps monitoring to further optimize SAP operations, including the following:

• Predictive Alerting and Adaptive Thresholds – Using machine learning to identify potential failures before they occur.
• Advanced RCA with IT Service Intelligence (ITSI) – Leveraging Splunk’s ITSI tool to visualize failure points in real time.
• End-to-End Business and IT Integration – Connecting technical monitoring with business processes for faster issue resolution.
• Expanding Agile DevOps Practices – Embedding AI-driven monitoring into development cycles to improve incident prevention.

Nasse relayed these key lessons learned from Keurig Dr. Pepper’s SAP Monitoring overhaul:

• Unifying SAP monitoring improves operational visibility, reduces downtime, and enhances RCA.
• AI-driven alerting and predictive analytics help eliminate false positives and optimize SAP maintenance.
• Real-time dashboards and automation significantly enhance SAP system efficiency and IT response times.
• Business and IT alignment through iDocs tracking and order monitoring improves customer service and order fulfillment.
• Migrating SAP systems to the cloud accelerates digital transformation, but effective monitoring is critical for success.

What this means for SAPinsiders

Avoid potential gotchas with Splunk. Integrating PowerConnect for SAP with Splunk can be complex, especially in multi-cloud or hybrid environments with customized SAP deployments. Misconfigurations can lead to incomplete data ingestion, incorrect alert thresholds, or blind spots in monitoring, leading to false positives or undetected critical incidents. Companies should conduct a phased rollout with thorough testing and tuning of alert settings to ensure accurate monitoring. Also, Splunk’s pricing model is based on ingested data volume, which can quickly escalate for enterprises processing large SAP log files and high-volume transactions. Without careful data management, companies may face exponential cost increases, reducing ROI. So companies should use indexing strategies to filter and prioritize logs, only storing critical data in Splunk, explore alternative pricing models (e.g., ingest vs. workload-based pricing), and leverage data lifecycle policies to offload older logs to cheaper storage solutions. And bear in mind that Splunk is not an SAP-native tool, meaning it lacks built-in functional understanding of SAP-specific modules like FI-CO, MM, SD, or S/4HANA workflows. Misinterpretation of SAP transaction logs can lead to false insights, incorrect root cause analysis, and misleading business decisions. So companies should ensure PowerConnect is properly configured to translate SAP logs correctly. Train IT and SAP support teams on Splunk’s SAP dashboards and data correlation techniques, and leverage SAP Business Technology Platform (BTP) integration for more contextual insights.

How to assess SAP system health. For a holistic view of SAP system health, operators should track Dialog Response Time – Ensures user transactions are processed efficiently; Work Process Utilization – Monitors dialog, update, batch, and spool processes; Database Performance – Tracks SAP HANA, Oracle, or SQL Server query execution times; Queue and Background Jobs – Avoids job failures that impact business operations; Memory and CPU Utilization – Prevents bottlenecks due to excessive memory consumption; and Network Latency – Ensures low-latency connectivity for SAP Fiori and S/4HANA cloud apps. This approach surfaces performance degradation early, ensuring smooth SAP operations.

AI raises the bar in SAP system monitoring. Leaders in roles like Nasse’s should use AI-driven anomaly detection to monitor system behavior and detect unusual patterns before failures occur. They should leverage ML models to analyze historical SAP performance data and predict potential bottlenecks (e.g., database slowdowns, memory leaks, high CPU usage), and integrate with SAP HANA predictive analytics to forecast system workload spikes during critical business cycles (e.g., month-end closing, peak sales periods). SAP customers should use ML algorithms to correlate system events, logs, and alerts across multiple SAP modules and infrastructure layers, and employ NLP (Natural Language Processing) models to analyze error messages and system logs for faster problem resolution. And like Nasse did, companies should leverage AI-powered log analytics tools (e.g., Splunk, Dynatrace, New Relic, SAP Focused Run) to automate RCA. AI can also help with issue resolution. Companies should consider implementing AI-powered self-healing mechanisms that detect system failures and trigger automated remediation scripts. AI chatbots and virtual assistants (e.g., SAP CoPilot, Microsoft Copilot, ServiceNow AI) can also provide automated troubleshooting guidance to SAP users. And intelligent RPA (Robotic Process Automation) bots can handle routine SAP error resolutions, background job restarts, and performance tuning tasks.