Detecting failed processes using fault signatures

A strategy is presented for automatically identifying processes that have failed. A determination is made of the types of data that need to be collected, and circumstances under which the approach is likely to be useful. The strategy is applied to generate signatures for three different types of workloads, and several different resources. A typical failure is injected into the process, and the associated signatures are presented for the same workloads and resources. A measure as defined that is used to determine whether or not a signature is likely to be indicative of a faulty process