Анализ логов и кластеризация ошибок

★ 6.6 · devops

log-error-digest is a Claude Code skill that analyzes log files to surface error patterns, frequency statistics, and time distribution reports through automated clustering. It runs `scripts/analyze_logs.py` with options to filter by log level (`--level`), time range (`--since`/`--until`), and the number of top clusters to display (`--top`). Three log formats are supported — JSON, syslog (RFC 3164), and Nginx — with automatic detection; output goes to the terminal or to a structured JSON file via `--output`. Error clustering normalizes dynamic message parts such as IPs, UUIDs, and numbers so that related failures are grouped and root causes become visible. Ideal for developers and DevOps engineers who need to debug a large log file, pinpoint peak error periods, or produce a report ready for integration with monitoring systems.