US 12,007,829 B2
Extended dynamic intelligent log analysis tool
Nirjar Gandhi, New Delhi (IN); Anviti Srivastava, Gurugam (IN); Sudhir Verma, Gurgaon (IN); Martin Adam, Prisnotice (CZ); and Vitezslav Visek, Brno (CZ)
Assigned to SAP SE, Walldorf (DE)
Filed by SAP SE, Walldorf (DE)
Filed on Oct. 31, 2022, as Appl. No. 17/977,683.
Claims priority of provisional application 63/419,908, filed on Oct. 27, 2022.
Prior Publication US 2024/0143430 A1, May 2, 2024
Int. Cl. G06F 11/30 (2006.01); G06F 11/07 (2006.01)
CPC G06F 11/0778 (2013.01) [G06F 11/0709 (2013.01); G06F 11/0781 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A computer implemented method comprising:
obtaining a log file comprising a plurality of log entries, each log entry comprising an error message and a timestamp;
identifying communities of errors, where identifying communities of errors comprises:
creating an error type for each unique error message in the plurality of log entries;
dividing the log file into a plurality of sessions, each session representing a predetermined period of time;
generating a graph, wherein each error type is plotted as a node in the graph;
determining, for a plurality of node pairs, a number of sessions in which both nodes of each node pair occur;
plotting edges between each node pair of the plurality of node pairs;
assigning a weight to each edge based on the determined number of sessions; and
performing a community detection algorithm on the graph to identify communities of errors; and
identifying anomalous sessions by identifying sessions comprising errors from more than one community.