What error patterns do you use in your Java services?

[openguardrail]

We designed this framework around grouped error codes with range validation, HTTP status mapping, and recoverability flags.

Curious to hear from the community:

• How do you structure errors in your services today?
• Do you find code range grouping useful for preventing collisions across teams?
• What's missing that would make this more useful for your use case?

Any feedback is welcome.