Skip to content

fix(strix): scan-start jitter plus gpt-5.6-luna fallback migration #1321

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

Closed
seonghobae wants to merge 2 commits into from
Closed

fix(strix): scan-start jitter plus gpt-5.6-luna fallback migration #1321

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
ce61587
fix(strix): de-synchronize concurrent PR scans with run-id jitter
seonghobae Aug 25, 2026
98733c7
fix(strix): complete the gpt-5.6-luna fallback migration pinned by po…
seonghobae Aug 25, 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
21 changes: 18 additions & 3 deletions .github/workflows/strix.yml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -453,7 +453,7 @@ jobs:
- name: Gate Strix secrets
id: gate
env:
STRIX_MODEL: ${{ github.event.client_payload.strix_llm || (steps.target_visibility.outputs.is_private == 'false' && 'nvidia_nim/nvidia/nemotron-3-super-120b-a12b' || 'gpt-5.4') }}
STRIX_MODEL: ${{ github.event.client_payload.strix_llm || (steps.target_visibility.outputs.is_private == 'false' && 'nvidia_nim/nvidia/nemotron-3-super-120b-a12b' || 'gpt-5.6-luna') }}
STRIX_MODEL_REQUESTED: ${{ github.event.client_payload.strix_llm || '' }}
STRIX_OPENAI_API_KEY: ${{ secrets.STRIX_OPENAI_API_KEY || secrets.OPENAI_API_KEY }}
STRIX_OPENROUTER_API_KEY: ${{ secrets.OPENROUTER_API_KEY }}
Expand All @@ -464,7 +464,7 @@ jobs:
run: |
strix_model="$(printf '%s' "$STRIX_MODEL" | sed 's/^[[:space:]]*//;s/[[:space:]]*$//')"
if [ -z "$STRIX_MODEL_REQUESTED" ] && [ "$strix_model" = "nvidia_nim/nvidia/nemotron-3-super-120b-a12b" ] && [ -z "${STRIX_NVIDIA_NIM_API_KEY:-}" ]; then
strix_model="gpt-5.4"
strix_model="gpt-5.6-luna"
fi
echo "strix_model=$strix_model" >> "$GITHUB_OUTPUT"
case "$strix_model" in
Expand Down Expand Up @@ -834,7 +834,7 @@ jobs:
STRIX_LLM_MAX_RETRIES: 1
STRIX_TRANSIENT_RETRY_PER_MODEL: 2
STRIX_TRANSIENT_RETRY_BACKOFF_SECONDS: 60
STRIX_FALLBACK_MODELS: ${{ steps.gate.outputs.provider_mode == 'github_models' && 'openai-direct/gpt-5.4' || steps.gate.outputs.provider_mode == 'openai_direct' && 'openai-direct/gpt-5.4' || steps.gate.outputs.provider_mode == 'openrouter' && 'openai-direct/gpt-5.4' || steps.gate.outputs.provider_mode == 'nvidia_nim' && 'nvidia_nim/nvidia/llama-3.3-nemotron-super-49b-v1.5 openai-direct/gpt-5.4' || '' }}
STRIX_FALLBACK_MODELS: ${{ steps.gate.outputs.provider_mode == 'github_models' && 'openai-direct/gpt-5.6-luna' || steps.gate.outputs.provider_mode == 'openai_direct' && 'openai-direct/gpt-5.6-luna' || steps.gate.outputs.provider_mode == 'openrouter' && 'openai-direct/gpt-5.6-luna' || steps.gate.outputs.provider_mode == 'nvidia_nim' && 'nvidia_nim/nvidia/llama-3.3-nemotron-super-49b-v1.5 openai-direct/gpt-5.6-luna' || '' }}

@devin-ai-integration devin-ai-integration Bot Aug 25, 2026

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

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

🟡 Model rename breaks pinned Strix workflow tests

Every gpt-5.4 string in the workflow became gpt-5.6-luna, but scripts/ci/test_strix_quick_gate.sh:306 and :363-365 still assert the workflow contains gpt-5.4, as does scripts/ci/strix_required_workflow_smoke.sh:171-175. Those assertions now fail and exit 1, so the required Strix quality check breaks.

Prompt for agents 
The workflow strix.yml was migrated so that all model references changed from gpt-5.4 to gpt-5.6-luna (default at line 456, gate fallback at line 467, STRIX_FALLBACK_MODELS at line 837). The workflow no longer contains any gpt-5.4 string. However the pinned contract tests still require gpt-5.4 and were not updated, so they now fail: scripts/ci/test_strix_quick_gate.sh lines 306, 363, 364, 365 assert_file_contains the old gpt-5.4 strings, and scripts/ci/strix_required_workflow_smoke.sh lines 171-175 assert_file_contains_either the nvidia_nim fallback with gpt-5.4. Update these assertions to the new gpt-5.6-luna strings (matching the exact new workflow content) so the Strix Changed Path Quality CI check and the required-path smoke self-test pass. Note per AGENTS.md that contract tests pin these workflow strings and must be kept in sync.
Open in Devin Review

Was this helpful? React with 👍 or 👎 to provide feedback.

STRIX_GITHUB_MODELS_API_BASE_FILE: ${{ env.STRIX_GITHUB_MODELS_API_BASE_FILE }}
STRIX_GITHUB_MODELS_KEY_FILE: ${{ env.STRIX_GITHUB_MODELS_KEY_FILE }}
STRIX_OPENAI_FALLBACK_KEY_FILE: ${{ env.STRIX_OPENAI_FALLBACK_KEY_FILE }}
Expand All @@ -855,6 +855,8 @@ jobs:
PR_BASE_SHA: ${{ github.event_name == 'pull_request_target' && github.event.pull_request.base.sha || github.event.client_payload.pr_base_sha }}
PR_HEAD_SHA: ${{ github.event_name == 'pull_request_target' && github.event.pull_request.head.sha || github.event.client_payload.pr_head_sha }}
IS_PR_EVIDENCE_RUN: ${{ (github.event_name == 'pull_request_target' || github.event.client_payload.pr_number != '') && 'true' || 'false' }}
github_run_id: ${{ github.run_id }}
github_run_attempt: ${{ github.run_attempt }}
run: |
budget_suffix="TIME""OUT"
process_budget_seconds="5400"
Expand All @@ -871,6 +873,19 @@ jobs:
# could not complete a scan. Provider failure is typed infrastructure
# evidence, but remains non-passing because no authoritative complete
# vulnerability result exists.

# De-synchronize concurrent PR scans. Several pull requests often
# start within the same minute; without a spread they all hit the
# same NVIDIA NIM / OpenAI capacity window at once and exhaust it
# with 429s before any retry ladder can recover. A run-id-derived
# jitter stays deterministic per run while bounding the delay well
# inside the process budget.
scan_stagger_seconds=$(( (github_run_id % 8) * 15 + (github_run_attempt - 1) * 30 ))

@devin-ai-integration devin-ai-integration Bot Aug 25, 2026

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

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

📝 Info: Stagger plus gate stays under the step timeout

Worst-case stagger is 165s (run_id%8==7, attempt 3). Added to the gate's 5700s total timeout that is 5865s, under the 100-minute (6000s) step timeout, leaving a ~135s margin. Bounded but tight.

Open in Devin Review

Was this helpful? React with 👍 or 👎 to provide feedback.

@devin-ai-integration devin-ai-integration Bot Aug 25, 2026

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

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

📝 Info: run_id % 8 bucketing can still collide

The stagger uses github_run_id % 8. Run ids are allocated near-sequentially, so concurrently queued PRs landing on the same residue get identical delays and still collide on the shared provider window. The spread is reduced, not guaranteed.

Open in Devin Review

Was this helpful? React with 👍 or 👎 to provide feedback.

if [ "$scan_stagger_seconds" -gt 0 ]; then
echo "Staggering Strix scan start by ${scan_stagger_seconds}s to avoid provider-capacity collisions across concurrent PR scans." >&2
sleep "$scan_stagger_seconds"
fi
Comment on lines +883 to +887

@devin-ai-integration devin-ai-integration Bot Aug 25, 2026

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

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

📝 Info: Zero-valued arithmetic expansion will not trip set -e

When github_run_id % 8 == 0 and attempt is 1, scan_stagger_seconds=$(( ... )) evaluates to 0. Arithmetic expansion $(( )) (unlike (( )) or let) never returns non-zero from its value, so the assignment succeeds under the shell's set -e. The -gt 0 guard then correctly skips the sleep.

Open in Devin Review

Was this helpful? React with 👍 or 👎 to provide feedback.


strix_run_log="$RUNNER_TEMP/strix_gate_console.log"
strix_rc=0
set +e
Expand Down
Loading