Our system already tracks many errors — most are expected and harmless (e.g. user interruptions, odd input, etc.). However, not all errors are equal, and treating them the same can overwhelm teams with noise.
Proposal: Smarter filtering for error notifications — for example:
- Prioritize repeated or high-frequency errors
- Flag issues affecting multiple users
- Highlight errors tied to key workflows or modules
