DETECTING NETWORK-APPLICATION SERVICE FAILURES

A computer-implemented process for detecting service failures of a network application under test involves collecting raw network-application availability monitoring (NAAM) data. The NAAM data can include datapoints, each which specifies the network application under test, a respective test site; a time corresponding to a respective probe, and an indication whether the respective probe resulted in a success or a failure. The raw NAAM data is filtered to remove datapoints indicating failures associated with causes other than a service failure of the NAUT. The filtered NAAM data is analyzed to detect a service failure of the NAUT.