Fixed
- (connectivity) The two L4 presets fit the smallest L4 instance — qwen3-4b-instruct and qwen3-8b-fp8 request 2 vCPU / 10 GiB (were 4 / 16Gi), limits 4 / 12Gi; verify-serving-slice holds the requests within what a g6.xlarge leaves a predictor (giantswarm/agent-platform#502) in #503 by @teemow
Full Changelog: https://github.com/giantswarm/agent-platform/compare/v4.28.8...v4.28.9