Repository Issue Activity (beta)

sgl-project/sglang

Current issue state, recent activity, and per-issue timelines from the indexed issue data.

Open Issues
704
New in 7 Days
40
Closed in 7 Days
28
Average Open Age
36 days
Stale 30+ Days
349
Stale 90+ Days
26
Last 2 Weeks
DateOpenedClosedCommentsEventsOpen Backlog
2026-07-1000000
2026-07-0900000
2026-07-0800000
2026-07-0700000
2026-07-06128387830
2026-07-05106194615
2026-07-0468224611
2026-07-03126357222
2026-07-0218124710837
2026-07-01149359232
2026-06-30147399436
2026-06-2915116315743
2026-06-28676110245
2026-06-277164112523
This Week

Opened: 28

Closed: 22

Comments: 79

Events: 170

Top Labels
inactive (2832)
high priority (405)
good first issue (248)
bug (215)
help wanted (184)
deepseek (91)
enhancement (68)
npu (64)
Issue Explorer
IssueAuthorStateLabelsCommentsReactionsUpdated

#30294 [Bug] Decode running batch pinned at `max_running_requests - 1` under Chunked Prefill

Opened 4 days ago
ITcarrot
open
No labels
004 days ago

#26340 CUDA Coredump Tracker

Opened 2 months ago
hnyls2002
open
No labels
18904 days ago

#17050 [Tracking] CI Test Failures and Fixes

Opened 6 months ago
alisonshao
open
No labels
13104 days ago

#29488 [Feature] Support DSpark Speculative Decoding for DeepSeek V4

Opened 13 days ago
Fridge003
closed - completed
No labels
474 days ago

#30200 [Performance] PR #29223 regresses Llama 3.3 FP8 TP8 H100 throughput by 3.2%

Opened 4 days ago
nvpohanh
open
No labels
104 days ago

#9319 [Feature] Dynamic Speculative Decoding

Opened 11 months ago
liangjason87
closed - completed
inactive
544 days ago

#29562 [Bug] GLM-5.2-NVFP4 error on pro6000

Opened 12 days ago
dengni-10000
open
No labels
1204 days ago

#30168 [Bug] Qwen3.6/GDN LoRA: `get_hidden_dim` not implemented for `in_proj_qkv`/`in_proj_z` — crashes server at startup

Opened 5 days ago
drodriguezhdez
open
No labels
304 days ago

#30158 [Bug] Compatibility issues with PP and Hicache causing errors

Opened 5 days ago
enternal111
open
No labels
804 days ago

#27356 [Feature] Add KVarN KV-cache quantization support for long-context reasoning

Opened 1 month ago
huxiang088
open
No labels
314 days ago

#30263 [Feature] Per-request opt-out of speculative decoding (RFC: disable_speculative_decoding sampling param)

Opened 4 days ago
anencore94
open
No labels
004 days ago

#30262 [Feature] Complete command-line configuration for multi-node and PD separation scenarios

Opened 4 days ago
linfordWu
open
No labels
004 days ago

#30057 [Bug] HiCache L3 partial prefetch can restore DSA/NSA KV pages without matching INDEXER sidecar pages

Opened 7 days ago
liaol
open
No labels
004 days ago

#21210 [Bug] EAGLE draft_forward leaks padded state across decode steps under DP attention, causing negative-dimension crash

Opened 4 months ago
lviy
closed - completed
inactive
204 days ago

#30239 [Bug] DeepSeek V4 loading fails with RunAI Model Streamer during FP8 `wo_a` dequantization

Opened 4 days ago
Broduker
open
No labels
004 days ago

#30245 ROCm consumer GPU (gfx1100/RDNA3): fused-MoE blocked by upstream triton AMD backend bug — request configs, allowlist fix, and mitigation

Opened 4 days ago
AIwork4me
open
No labels
004 days ago

#22511 [Bug] sglang v0.5.10 2 * H100 deploy glm5 report health check failed. server could not get a response from detokenizer for last 20 seconds

Opened 3 months ago
wangxiaoyang-dev
open
high priority
2444 days ago

#29599 [npu] Does Latest SGLang Support DeepSeek V4 on Ascend NPU now?

Opened 11 days ago
razIove
open
No labels
214 days ago

#30242 [Bug] Deepseek V4 NVFP4 Disagg Decode node failure

Opened 4 days ago
shyeh25
open
No labels
104 days ago

#30233 [Bug] PD disaggregation: aborted prefill requests (input too long) cause decode to generate garbage from incomplete KV cache

Opened 4 days ago
kevin-14
open
No labels
004 days ago

#27368 [Bug] NUMA: scheduler subprocess dies (exit 1, no traceback) when `numactl --membind` fails under a constrained container

Opened 1 month ago
chengcuiping
closed - completed
No labels
104 days ago

#30224 [Feature] Support LongCat2.0 for Ascend NPU

Opened 4 days ago
LinyuanLi0046
open
No labels
004 days ago

#28618 [RFC] Add SM89/L20 support for DeepSeek-V4-Flash-FP8

Opened 22 days ago
xesdiny
open
No labels
304 days ago

#25371 [RFC][Refactor] `mem_cache` pool / allocator restructure

Opened 2 months ago
alphabetc1
open
No labels
024 days ago

#30084 The deepseekv4 model experienced a performance degradation in ttft when subjected to pd separation.

Opened 6 days ago
BJWang-ant
open
No labels
104 days ago

Rows per page:

1–25 of 6,410