The Monitor System for the LHCb on-line farm

The aim of the LHCb on-line farm Monitor System is to keep under control all the working indicators which are relevant for the farm operation, and to set the appropriate alarms whenever an error or a critical condition comes up. Since the most stressing tasks of the farm are the data transfer and processing, relevant indicators includes the CPU and the memory load of the system, the network interface and the TCP/IP stack parameters, the rates of the interrupts raised by the network interface card and the detailed status of the running processes. The monitoring of computers’ physical conditions (temperatures, fan speeds and motherboard voltages) are the subject of a separate technical note, since they are accessed in a different way, by using the IPMI protocol.