What’s Changed
- chore(deps): update go toolchain directive to v1.26.3 by @renovate[bot] in https://github.com/giantswarm/klausctl/pull/217
Full Changelog: https://github.com/giantswarm/klausctl/compare/v0.0.78...v0.0.79
Changes and Releases
Updates on Giant Swarm workload cluster releases, apps, UI improvements and documentation changes.
What’s Changed
- fix(deps): update module github.com/giantswarm/mcp-oauth to v0.2.119 by @renovate[bot] in https://github.com/giantswarm/muster/pull/645
Full Changelog: https://github.com/giantswarm/muster/compare/v0.1.157...v0.1.158
What’s Changed
- fix(deps): update module github.com/mark3labs/mcp-go to v0.52.0 by @renovate[bot] in https://github.com/giantswarm/klausctl/pull/219
Full Changelog: https://github.com/giantswarm/klausctl/compare/v0.0.77...v0.0.78
What’s Changed
- chore(deps): update dependency architect to v8 by @renovate[bot] in https://github.com/giantswarm/klausctl/pull/220
Full Changelog: https://github.com/giantswarm/klausctl/compare/v0.0.76...v0.0.77
What’s Changed
- fix(deps): update module github.com/giantswarm/klaus-oci to v0.0.24 by @renovate[bot] in https://github.com/giantswarm/klausctl/pull/218
Full Changelog: https://github.com/giantswarm/klausctl/compare/v0.0.75...v0.0.76
Fixed
- Stabilise the AI chat transport across React renders.
useChatSetupconstructed a freshAssistantChatTransporton every render and handed it touseChatRuntime; when its identity changed mid-stream (e.g. as a side effect of a state update triggered by streamedreasoning-delta/tool-input-deltaevents), the runtime tore down the in-flight chat fetch withTypeError: network error. Envoy logged the symptom asresponse_flags: DC(downstream remote disconnect) against a healthy Backstage upstream while the SSE stream summary wassawFinish=false, last eventtool-input-delta, surfacing in the UI as a “Network error” banner even though no real network outage occurred. The transport is now memoised on the resolved API URL and the stablegetHeaders/debugFetchcallbacks, so a single transport lives for the component’s lifetime. - AI chat instrumentation: observe the caller’s
AbortSignaland log abort events with the reason inline (ABORT signaled by client at <ms> -- reason: <name>: <message>). Stream-outcome lines now annotate aborted streams with[client-aborted at Nms reason="..."]so a “Network error” banner can be classified as a deliberate client cancel (transport rebuild, unmount, manual stop) versus a real proxy / network failure (no abort signal fired, raw stream error). Read errors that fired after the caller aborted are now logged asconsole.warn(“cancelled by client AbortSignal”) rather thanconsole.error(“STREAM READ FAILED”), reserving the latter for the genuine pre-finish, non-aborted failure mode.
See ./docs/releases/v0.129.3-changelog.md for more information.
- Stabilise the AI chat transport across React renders.
Fixed
- AI chat instrumentation: render the SSE stream summary inline in the log message (
bytes=... events=... sawFinish=... lastEventType=...) instead of attaching it as a trailing object argument, so it stays readable in browser-console consumers that flatten the args array (devtools-snapshotters, log shippers, the Cursor IDE browser, etc.) rather than collapsing to “[object Object]”. Also classify a stream read error that arrives after afinishevent was already parsed as a post-completion teardown (console.warn, message already committed) instead of a network failure (console.error); this matches the actual semantics of the AI SDK aborting the underlying fetch once it has finished consuming the stream. See ./docs/releases/v0.129.2-changelog.md for more information.
- AI chat instrumentation: render the SSE stream summary inline in the log message (
Fixed
- AI chat: always log network-level diagnostics (request URL, response status / headers, fetch errors, SSE stream lifecycle including premature termination) to the browser console so a “Network error” banner can be triaged from a user report without requiring the
ai-chat-verbose-debuggingfeature flag. Verbose payload-level logging (messages, system prompt, tool schemas, per-event SSE detail) remains gated on the feature flag in non-production builds. See ./docs/releases/v0.129.1-changelog.md for more information.
- AI chat: always log network-level diagnostics (request URL, response status / headers, fetch errors, SSE stream lifecycle including premature termination) to the browser console so a “Network error” banner can be triaged from a user report without requiring the
Added
- Make AI chat sampling parameters configurable per installation. The plugin previously called
streamText()withouttemperature,topP,topK,seed,minP, ormaxOutputTokens, so the server’s defaults applied – which for vLLM meanstemperature=1.0, top_p=1.0, top_k=-1, seed=null. That is far too loose for a tool-using agent backed by a reasoning model and was the dominant cause of token-cost variance in production agent loops (same prompt, fresh chat, observed total-token spread of 22k / 607k / 22k across three runs against the same Qwen3 endpoint). Config now accepts anaiChat.samplingblock withtemperature,topP,topK,minP,seed, andmaxOutputTokens; all fields are optional and default behaviour with nosampling:block is unchanged.temperature,topP,topK,seed, andmaxOutputTokensare forwarded through the AI SDK to every provider that supports them.minPis spliced into the request body via the OpenAI-compatible provider’stransformRequestBodyhook, since vLLM accepts it as a top-level field but it is not part of the AI SDK call settings. Theai-chat-backendREADME documents recommended values per model family (Qwen3 thinking/non-thinking, Qwen3-Coder, GPT-4 / GPT-4o, Anthropic Claude). See ./docs/releases/v0.129.0-changelog.md for more information.
- Make AI chat sampling parameters configurable per installation. The plugin previously called
What’s Changed
- fix(aggregator): expose workflows as workflow_* only, drop core_action_* advertisement by @teemow in https://github.com/giantswarm/muster/pull/646
Full Changelog: https://github.com/giantswarm/muster/compare/v0.1.156...v0.1.157