In order to face the high market competiveness, the Power-OMproject (http://power-om.eu/) aims at implementing a proactiveapproach for improving the machine tool performances. Forimplementing a proactive approach that helps monitoring machinetool performances, this paper presents a technical architecturewith two levels: the local and the remote one. In the local level,condition based maintenance strategy is implemented and realtime data is used for monitoring the local health of a machine. Afirst originality is to use the current analysis for assessing thehealth status of the machine. In the remote level, offline data isstored in an eMaintenance platform, which allows providing afleet dimension. This dimension allows to benefit of more dataand information allowing to make performances comparisonacross the fleet and along the time.This paper presents the advantages of the added-valuearchitecture. On one hand, there is the possibility to trackperformances to detect drifts locally and real-time based on thecurrent analysis, and on the other hand, to follow-up short andmid-term performances for deeper analysis of the fleetperformances in order to bring relevant information for decisionmakers.