A method, computer program product, and system for performance monitoring of a computer resource utilized by a user community. The computer resource is monitored, and sets of performance metrics characterizing a state of the computer resource are collected and analyzed to detect performance incidents. External information such as contextual information and user annotations are also collected and analyzed. If a performance incident is detected, its set of performance metrics and associated external information are used to search a data repository of past performance incidents for related incidents, in order to identify critical performance incidents, determine recommended solutions for the performance incidents, and determine performance forecasts for the computer resource.