The performance and efficiency of LEP depend on a multitude of factors, including particle beam characteristics, physics parameters, hardware settings, and environmental conditions. Often, these factors interact in unexpected ways and affect the machine performance. In January 1992, a project was started to create a unique logging system using an on-line database. One year's worth of data was to be kept on-line, which was estimated to be a total of 8 GByte. The systems concerned are of different nature (particle beam profile, power converter current, meteorological data, magnet temperature, lep-mode, …), require sampling at different frequencies (from seconds to several hours) and are of different sample size (from 10 MByte to 1 GByte per year). Major performance criteria included rapid logging of data for useful real-time monitoring of compound measurements, and rapid retrieval and correlation of large amounts of data for efficient off-line analysis. For the database design the NIAM methodology was used as well as some interesting techniques such as tagging the rows with timeslots instead of timestamps and row packing for storage minimisation. A complex structure of servers and clients takes care of data gathering, data logging and management of all real time measurement and logging requests. Several tools have been developed to make the data correlation transparent to non-database experts.