Algorithm for aggregating relevant log statements from distributed components, which appropriately describes an error condition

摘要

A method for aggregating relevant log statements from distributed components is provided. The method includes representing a plurality of error codes of an application as nodes in a graph, with each of the plurality of error codes connected to at least one other of the plurality of error codes by a causal relationship represented in the graph. The method includes determining causal relationships among a plurality of log statements from distributed components, based on correlating error codes in the plurality of log statements in accordance with causal relationship connections represented in the graph, and determining a subset of information from the plurality of log statements as one or more single composite events, each comprising a plurality of error codes, in causal relationship in accordance with the graph, based on the determining the causal relationships.