Search results

[BUG] metrics-driven-autoscaling guide builds an HPA on llamacpp:kv_cache_usage_ratio, which llama.cpp does not export

Status: Open.

#1373 In defilantech/LLMKube;

· Tanguille opened on Jul 30, 2026

[FEATURE] Operator-emitted Prometheus scrape annotations for inference pods (no prometheus-operator)

Status: Open.

#1367 In defilantech/LLMKube;

· brandonliuli opened on Jul 30, 2026

[BUG] revision restore silently falls back to base and force-pushes away the prior attempt (reviewed work destroyed)

bugSomething isn't workingSomething isn't working

Status: Open.

#1364 In defilantech/LLMKube;

· joryirving opened on Jul 30, 2026

[FEATURE] Operator-health row in llmkube-inference (reconcile rate, reconcile p95, ready duration)

Status: Open.

#1362 In defilantech/LLMKube;

· Tanguille opened on Jul 30, 2026

[FEATURE] SGLang dashboard: queue-wait, per-stage latency, new_token_ratio and an L3 HiCache row

Status: Open.

#1360 In defilantech/LLMKube;

· Tanguille opened on Jul 30, 2026

[FEATURE] llama.cpp runtime dashboard (the runtime with no board of its own)

Status: Open.

#1359 In defilantech/LLMKube;

· Tanguille opened on Jul 30, 2026

[BUG] gpu alert group is DCGM-only and default-on, so AMD clusters get silent no-op alerting

Status: Open.

#1356 In defilantech/LLMKube;

· Tanguille opened on Jul 30, 2026

[BUG] Unused per-InferenceService model-cache PVC provisioned for pvc:// (pre-staged) sources

Status: Open.

#1353 In defilantech/LLMKube;

· brandonliuli opened on Jul 30, 2026

[FEATURE] Coverage/mutation-delta gate for behavior changes lacking a fail-on-old-behavior test (#1332 follow-up)

area/foremanForeman: the agentic fleet orchestrator add-onForeman: the agentic fleet orchestrator add-on

enhancementNew feature or requestNew feature or request

Status: Open.

#1348 In defilantech/LLMKube;

· Defilan opened on Jul 29, 2026

[BUG] Coder gate cannot detect assertion-value churn that encodes buggy behavior as expected (#1332 follow-up)

area/foremanForeman: the agentic fleet orchestrator add-onForeman: the agentic fleet orchestrator add-on

bugSomething isn't workingSomething isn't working

Status: Open.

#1347 In defilantech/LLMKube;

· Defilan opened on Jul 29, 2026

[BUG] Coder gate cannot detect string-match-only tests for command-string/shell changes (#1332 follow-up)

area/foremanForeman: the agentic fleet orchestrator add-onForeman: the agentic fleet orchestrator add-on

bugSomething isn't workingSomething isn't working

Status: Open.

#1346 In defilantech/LLMKube;

· Defilan opened on Jul 29, 2026

[FEATURE] Record MCP server config + per-tool call histogram in the Foreman audit record (foreman.audit.v1)

enhancementNew feature or requestNew feature or request

Status: Open.

#1331 In defilantech/LLMKube;

· Defilan opened on Jul 28, 2026