Skip to content

ci: increase gpu timeout budgets #1721

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
key4ng merged 2 commits into from
Jun 13, 2026
Merged

ci: increase gpu timeout budgets #1721

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
b6104c5
ci: increase gpu timeout budgets
key4ng Jun 13, 2026
940d292
ci: tune sglang chat timeout budget
key4ng Jun 13, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
7 changes: 4 additions & 3 deletions .github/workflows/pr-test-rust.yml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -315,7 +315,7 @@ jobs:
benchmarks:
needs: build-wheel
runs-on: 4-gpu-h100
timeout-minutes: 30
timeout-minutes: 36
permissions:
contents: read
steps:
Expand Down Expand Up @@ -357,6 +357,7 @@ jobs:
env:
ROUTER_LOCAL_MODEL_PATH: /models
E2E_LOG_DIR: benchmark-logs
GENAI_BENCH_TEST_TIMEOUT: "480"
run: |
mkdir -p benchmark-logs
bash scripts/ci_killall_sglang.sh "nuke_gpus"
Expand Down Expand Up @@ -494,8 +495,8 @@ jobs:
matrix:
include:
- engine: sglang
timeout: 28
test_timeout: 20
timeout: 36
Comment thread
key4ng marked this conversation as resolved.
test_timeout: 28
Comment on lines +498 to +499

@chatgpt-codex-connector chatgpt-codex-connector Bot Jun 13, 2026

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

P2 Badge Apply the intended SGLang chat timeout budget

For the e2e-1gpu-chat (sglang) matrix entry, these two values are forwarded directly to the reusable workflow as the job timeout and the Run E2E tests step timeout (.github/workflows/e2e-gpu-job.yml:56 and :131). The change notes say this shard needs 40 minutes overall and 32 minutes for pytest after the slow SGLang startup plus rerun case, but this leaves it at 36/28, so any run in that 28–32 minute pytest window will still be cancelled before completing. Please set this entry to the intended 40/32 budget.

Useful? React with 👍 / 👎.

Comment thread
key4ng marked this conversation as resolved.
- engine: vllm
timeout: 24
test_timeout: 18
Expand Down
Loading