Fixed
- (deps) Update module github.com/giantswarm/klaus-oci to v0.0.75 in #214 by @renovate[bot]
Full Changelog: https://github.com/giantswarm/klaus-operator/compare/v0.0.162...v0.0.163
Changes and Releases
Updates on Giant Swarm workload cluster releases, apps, UI improvements and documentation changes.
Changed
- (deps) Update dependency @anthropic-ai/claude-code to v2.1.275 in #490 by @renovate[bot]
Full Changelog: https://github.com/giantswarm/klaus/compare/v0.1.21...v0.1.22
Fixed
- (deps) Update module github.com/giantswarm/klaus-oci to v0.0.75 in #350 by @renovate[bot]
Full Changelog: https://github.com/giantswarm/klausctl/compare/v0.4.7...v0.4.8
Fixed
- (agent-platform) The serve intent’s step shows the served model’s steps with the read after load_model, in a tab that is not focused too — no reload in #2447 by @teemow
Full Changelog: https://github.com/giantswarm/backstage/compare/v2.23.0...v2.23.1
Fixed
- (model-serving) VLLM’s cache is a subPath of the claim’s own — .vllm-cache at the claim root mounted at /mnt/vllm-cache on the runtime container, VLLM_CACHE_ROOT set in the same rule — never a path under the model’s directory in #542 by @teemow
Full Changelog: https://github.com/giantswarm/agent-platform/compare/v4.32.0...v4.32.1
Added
- (cluster-manager) The connectivity chart ships the PriorityClass the gpu-node-pool chart’s prewarm placeholder runs at — agent-platform-prewarm-placeholder (-1000, Never), rendered with the cluster-manager component, mirrored in the meta chart in #540 by @teemow
Full Changelog: https://github.com/giantswarm/agent-platform/compare/v4.31.0...v4.32.0
Added
- (model-serving) The cache outlives a pool teardown — the serving namespace is kept while the cache is on, the claim gets a fast gp3 StorageClass of its own by default, and vLLM’s compile cache lives on the claim (VLLM_CACHE_ROOT) in #538 by @teemow
Full Changelog: https://github.com/giantswarm/agent-platform/compare/v4.30.3...v4.31.0
Added
- (repositories) Reconcile now becomes Align now — the dialog says the run changes the repository on GitHub and CircleCI to its declared set-up and the company baseline as the person; the dry run shows the manager’s warning, the team’s opt-in and the planned changes per step; Check now without the opt-in in #2440 by @teemow
Full Changelog: https://github.com/giantswarm/backstage/compare/v2.22.0...v2.23.0
Added
- (agent-platform) The preset chosen under I want to serve is the pool’s serve intent — served through model-manager as the person once the serving stack is ready, the step shows the served model’s timeline, a reload keeps it, Remove clears it in #2439 by @teemow
Full Changelog: https://github.com/giantswarm/backstage/compare/v2.21.3...v2.22.0
Fixed
- (model-serving) Preset qwen3-8-27b declares the Hub’s weight size (25 GiB) and its target hardware; every preset’s weightsGiB is verified against the Hub in CI (#535) in #536 by @teemow
Full Changelog: https://github.com/giantswarm/agent-platform/compare/v4.30.2...v4.30.3