LogAssist: Assisting Log Analysis Through Log Summarization

Logs contain valuable information about the runtime behaviors of software systems. Thus, practitioners rely on logs for various tasks such as debugging, system comprehension, and anomaly detection. However, are difficult to analyze due their unstructured nature large size. In this paper, we propose a novel approach called LogAssist that assists with log analysis. provides an organized concise view by first grouping into event sequences (i.e., workflows), which better illustrate execution paths. Then, compresses events in workflows hiding consecutive applying n-gram modeling identify common sequences. We evaluated generated one enterprise two open source find can reduce number need investigate up 99 percent. Through user study 19 participants, assist reducing time required analysis average 40 The participants also rated 4.53 out 5 improving experiences performing Finally, document our lessons learned from developing adopting practice. believe reported may lay basis future interactive exploration logs.