Grayalert: new tool to detect not yet seen errors

Dear Graylog Community,

I’ve been using Graylog for years and now developed a tool to detect new errors (especially useful after the release of a new version of a microservice or so).

I’m pleased to introduce GrayAlert, a new tool designed to complement your Graylog setup. GrayAlert focuses on enhancing error log monitoring by notifying you of new and unique error logs, helping to reduce the volume of redundant log entries.

Features of GrayAlert include:

  • anomaly detection using suffix trees/PATRICIA tries.
  • Integration with Graylog.
  • Data storage and a web interface for message review.
  • Notification support for Microsoft Teams, with potential for additional platforms.

I’m open to contributions for further development, including better anomaly detection, support for more logging platforms, more notification options, and the generation of daily summary reports.

For more information and to contribute, please visit our GitHub page: GrayAlert GitHub

Thank you for your interest and support.


I wonder, how much interest is in general in anomaly detection in log messages? Error messages are indeed interesting, but frequently there are too many of them…
Has anyone tried to apply any of the anomaly detection techniques from GitHub - logpai/awesome-log-analysis: A list of awesome research on log analysis, anomaly detection, fault localization, and AIOps?

1 Like