Event logs are the first place where to find usefulinformation about application failures. Event logs are availableat different system levels, such as application, middlewareand operating system. In this paper we analyze the failurereporting capability of event logs collected at different levels ofan industrial system in the Air Traffic Control (ATC) domain.The study is based on a data set of 3,159 failures inducedin the system by means of software fault injection. Resultsindicate that the reporting ability of event logs collected at agiven level is strongly affected by the type of failure observedat runtime. For example, even if operating system logs catchalmost all application crashes, they are strongly ineffective inface of silent and erratic failures in the considered system.
What Logs Should You Look at When an Application Fails? Insights from an Industrial Case Study
Antonio Pecchia
2014-01-01
Abstract
Event logs are the first place where to find usefulinformation about application failures. Event logs are availableat different system levels, such as application, middlewareand operating system. In this paper we analyze the failurereporting capability of event logs collected at different levels ofan industrial system in the Air Traffic Control (ATC) domain.The study is based on a data set of 3,159 failures inducedin the system by means of software fault injection. Resultsindicate that the reporting ability of event logs collected at agiven level is strongly affected by the type of failure observedat runtime. For example, even if operating system logs catchalmost all application crashes, they are strongly ineffective inface of silent and erratic failures in the considered system.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
