docs(e2e): spike scheduled server nightly - #6
Closed
localai-org-maint-bot wants to merge 1 commit into
Closed
Conversation
localai-org-maint-bot
force-pushed
the
codex/serve-e2e-nightly-spike
branch
30 times, most recently
from
July 27, 2026 12:04
37686da to
9ab554a
Compare
localai-org-maint-bot
force-pushed
the
codex/serve-e2e-nightly-spike
branch
26 times, most recently
from
July 30, 2026 03:03
90d9f40 to
5aae62f
Compare
Inventory the pinned vLLM entrypoint and V1 e2e CI topology, define a fail-closed CPU-presubmit versus scheduled-GB10 split, and decompose the implementation into independently gateable leaves. FOLLOWING_AGENTS_PROTOCOL Assisted-by: Codex:GPT-5 [Codex]
localai-org-maint-bot
force-pushed
the
codex/serve-e2e-nightly-spike
branch
from
August 3, 2026 03:17
5aae62f to
d267782
Compare
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Outcome
Spikes
SERVE-E2E-NIGHTLYwithout requiring GPU work. The design inventories pinned vLLMe24d1b24entrypoint/V1-e2e CI, records the current local orchestration gap, and defines a fail-closed CPU-presubmit versus scheduled-GB10 split.Design
N1/N2/N3; GB10 workflow and release-totality leavesN4/N5remain separately claimableNo source, test, workflow, checkpoint, GPU, or benchmark was changed or run.
docs/BENCHMARKS.mdrecordsNOT APPLICABLE.Verification
python3 scripts/check-agent-record.pypython3 tests/scripts/test_agent_record.py(13 tests)python3 scripts/check-doc-checkpoint.pypython3 tests/scripts/test_doc_checkpoint.py(5 tests)python3 scripts/check-device-leakage.py(DSR 32, baseline 32)git diff --check HEAD^ HEADAll pass on a GPU-less host.