A method is described of managing service events in a distributed computing system. The distributed computing system comprises a plurality of computing nodes able to perform a service using a service process. The method takes place at one of the computing nodes. A service event is received or created. This service event is identified by a combination of a node identifier, a time element, and a local counter value. The local counter value represents a number of service events performed by a service process for a user since a last reset. The identified service event is then stored in a service process database according to node identifier and local counter values. The service process database is used to manage service events in the distributed system. Service events are removed from the service process database when no longer valid using the time element.