Coop Pronto AG uses PRTG to monitor the availability and technical condition of its IT infrastructure. The existing setup provides reliable device monitoring but only limited visibility into gradual anomalies, cross-system dependencies, transaction-level disruptions, and their business impact. The thesis therefore examines whether AI-enhanced observability can combine technical and business data, broaden the current monitoring approach, and support earlier, more contextual disruption management across its entire IT environment.

The study combines a literature review, expert interviews, a current-state analysis of the existing IT and monitoring environment, and the iterative implementation and evaluation of a Dynatrace proof of concept. Controlled test scenarios and observed incidents were used to assess infrastructure monitoring, business-process visibility, anomaly detection, root cause support, and operational workflows. Datadog, Dynatrace, and Splunk were also compared through a document-based feature benchmark, a technical-fit assessment, and vendor quotations.

Dynatrace extended monitoring beyond device availability by combining telemetry, logs, business events, and store data. AI-supported anomaly detection identified technical and transactional anomalies, including gradual increases in transaction duration. Payment failures, POS errors, and communication issues became visible even when devices remained reachable, supporting earlier awareness and contextual incident analysis. However, Dynatrace provided no detection advantage over PRTG for conventional infrastructure outages. Dynatrace automatically detected and visualized relationships between selected OneAgent-monitored systems through Smartscape, but Dynatrace Intelligence did not identify the root cause in the tested incidents. Their dependencies were therefore modeled through predefined workflows, which supported cause assessment but required manual logic and maintenance. Dynatrace achieved the strongest functional coverage and technical fit and was the only platform validated through a proof of concept. Its cost was comparable to Datadog and slightly above Splunk Observability Cloud.