feat(health): AVX2 gate + recall_tier surfacing - #14
Merged
Conversation
- daimon-vec: embedder_supported() CPU gate; Embedder::new() now returns a clear error on non-AVX2 x86_64 instead of letting ort SIGILL at first inference (aarch64 unaffected) - daimon-mcp: /readyz and /health report recall_tier (hybrid | keyword | unhealthy); existing degradation path now catches the AVX2 case with a clear warning - daimon-indexer: embedder failure parks the indexer instead of crash-looping under compose/k8s restart policy (--once still errors); backfill later with daimon reindex - daimon-cli: health output gains recall_tier - install.sh: AVX2 preflight after stack-up (x86_64 only) with a pointer to the readyz recall_tier field - readme: deployment section documents the degradation contract Reported by @pizofreude in #2/#4.
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Addresses #2 and #4, reported by @pizofreude.
What changed
embedder_supported()CPU-feature gate.Embedder::new()now returns a clear error on non-AVX2 x86_64 instead of letting the ONNX runtime SIGILL at first inference, long after startup looked healthy. aarch64 (NEON path) is unaffected./readyzand/healthnow reportrecall_tier(hybrid | keyword | unhealthy). The existing graceful-degradation path catches the AVX2 case with a single clear warning, exactly as proposed in install.sh: preflight check for AVX2 before declaring the stack ready #2.--oncestill errors for CI use. Semantic backfill after moving to capable hardware:daimon reindex.daimon healthoutput gainsrecall_tier, computed the same way.recall_tierfield for later checks.Verification
cargo test --workspace: 37 passed, 0 failed (includes a newrecall_tierunit test)bash -n install.shclean; preflight is arch-guarded so arm64 hosts don't false-warnCloses #2, closes #4.