A method and system for improving a functioning of critical execution computing machines. A promoted critical execution computing machine determines one or more correction actions for reducing an impact of an execution issue on execution of critical work units by respective critical execution computing machines on the execution of the critical work units by the respective critical execution computing machines according to global correction information, local correction information, and peer correction information. The global correction information had been received from a scheduling computing machine and relates to the critical work units and to the critical execution computing machines. The local correction information relates to each corresponding critical work unit and to a status of the promoted critical execution computing machine. The peer correction information relates to the corresponding critical work units and to a corresponding status of the other critical execution computing machines.