Techniques are described for monitoring resource consumption, detecting faults, and predicting future faults with internal or customer-facing portal applications that engage data center operations for supporting co-location and, in some cases, interconnection by customers of a co-location provider. For example, an operational intelligence engine for portal applications deployed by a co-location provider stitches together transactions records and logs based on a unique identifier inserted into transactions performed by multiple system applications that execute the portal transactions through a lifecycle of requests and events for the portal and system applications.