Detecting and recovering from process failures

A service is used to process files. The processing of the files is performed by worker services that are assigned to process a portion of the files. Each worker service that is processing a portion of the files is assigned a unique identifier. Using the identifier information, the set of worker services currently active are monitored along with the work assigned to each process. When a worker server determines that a worker service has failed, the work assigned to the failed worker service can be automatically determined and a new worker service can be started to process that work. Any new worker service that is started is assigned a unique identifier, so the work assigned to it can be similarly tracked.