A tool idea for cutting app log costs

Sending every app or server log to an expensive live logging service can become costly. A lot of that volume may be repeated information logs, access records, health checks, and other low-value noise. The proposed approach is to send logs to different places based on how important they are.

Error, , and would stay in fast searchable storage, while repeated information logs and access records would go to a warehouse or summary view. Full raw logs would still be kept in cheap , so the team would not lose the original evidence. Common patterns and examples could also become context for .

A small browser-only analyzer has been built to test the idea.

Key points

  • Sending 100% of logs to live logging tools can create high bills.
  • Repeated information logs, access records, and health checks may not need premium storage.
  • Error, , and should stay easy to search quickly.
  • Full raw logs can still be in cheaper .
  • A safe trial should start with one service and clear rules for protected logs.
Read original