COMMITS
September 11, 2026
N
add guides for ROCm tiered prefix cache (native + lmcache, cpu + fs) (#2464)
nitsingh2 committed
R
docs(api): document requestHandler.parsers as a list (#2382)
Rishabh Sinha committed
B
feat(guides): add diffusion-serving text-to-speech guide (#2425)
Bob Tian committed
Y
L
docs(README): fix stale version badge and theme count mismatch (#2367)
Lily Sam committed
K
docs(observability): add batch-gateway dashboards and alerts to llm-d (#2390)
Kush Zingade committed
Y
docs: add graceful shutdown & request draining ops guide (#2444)
Yanshu committed
J
docs: fix ModelExpress base image component path (#2451)
jxj committed
M
Temporarily disabled scheduled execution of all OpenShit (IBM) workflows (#2477)
maugustosilva committed
J
proposal: add MetaX accelerator support (#2308)
Jifei Wang committed
J
docs: document GKE gateway guide prerequisites (#2448)
jxj committed
P
feat(workload-autoscaling): add AMD ROCm support for the WVA nightly (#2467)
Prateek Kumar committed
September 10, 2026
D
Expand and update the release matrix (#2471)
Diego Castan committed
P
ci(workload-autoscaling): add AMD ROCm WVA nightly workflows (#2469)
Prateek Kumar committed
B
updating fma+keda benchmark report (+ new report for Qwen3-14B) (#2459)
Braulio Dumba committed
D
tiered-prefix-cache: Add TPU multi-host support and restructure overlays (#2313)
dannawang0221 committed
September 9, 2026
Y
feat(predicted-latency-routing): add Intel XPU support (#2460)
Yao Qing committed
R
feat(snapshot): add GKE fast pod snapshot provider and vLLM wrapper (#2105)
Ryan R. Rosario committed
X
feat(modelexpress-p2p): add gke overlay for A3 Ultra (DRA/RoCE) (#2423)
Xia Hua committed
N
R
fix(guides): configure homogeneous TP=2 for SGLang PD disaggregation (#2442)
Rahul Gurnani committed
September 8, 2026
J
fix: inherit variables between Dockerfile stages (#2446)
jxj committed
J
fix(flow-control): validate tuning wizard inputs (#2455)
jxj committed
R
fix: correct 'retuned' to 'returned' typo in flow-control/README.md (closes #2303) (#2305)
RoomWithOutRoof committed
R
Add preferred backend guide to llm-d documentation (#2378)
Radhika Lakhtakia committed
J
fix: preserve healthcheck failure accounting (#2401)
jxj committed
J
fix: report unhashable guide YAML keys (#2406)
jxj committed
J
fix(ci): wait for flow-control request bursts (#2438)
jxj committed
W
Move GKE Node Selection From Base Into Provider Overlays (#2430)
weizhoublue committed
J
fix: handle malformed metrics summaries (#2450)
jxj committed