World of Logs Dataset

World of Logs (WoL) is a dataset comprising software logs and their associated context, including the natural language source text from which the logs are extracted and metadata such as project name, bug severity, creation date, and source URL. WoL allows researchers to quantify and examine sets of logs based on key features, enabling them to discover trends in log practices online. The WoL dataset is stored as a MongoDB dump, comprising 2,585,830 log messages and logging statements collected from the web by November 10, 2025. We have extracted software logs from 2,686,223 JIRA issues, 799,968 GitHub comments (i.e., issues and pull requests), and 2,759,000 SO questions, and are continuously updating our dataset. Based on WoL, we built LogSearch, a specialized log search engine, enabling developers to search, filter, and explore logs with their full contextual information. - https://search.logsearches.org/