Cross-request KV cache exposure in the GLM-5.3 Workers AI model, and our remediation with SGLang
Imported from official source
A configuration mismatch in GLM 5.3 on Workers AI triggered an SGLang bug that exposed KV cache pages across requests. Here is how we addressed the issue.
This version
- Version
- 1 of 1
- Recorded
- October 26, 2026 22:02
- Change
- Initial
- Content hash
9fba35c1c42e7ac5c01f028d3b6a71b0- All versions
- Revision history