Conversation
Consolidate the production-grade plan for user-managed multi-document chunk reuse into a single optimization record under ``docs/optimizations/``. The document captures positioning, SLA targets, connector capability boundary, staged roadmap (observability, DocSet management, hierarchical cache, accuracy safety net), data model extensions, CLI flags, test matrix, risks, rollback plan, and PR breakdown. No source code changes; later PRs land each stage behind independent feature flags.
Lay the metadata foundation for production observability and the later LFU+LRU hybrid eviction policy. - ChunkMeta gains ``access_count`` and ``last_access_time``; both default to zero / ``created_at`` so existing call sites are unaffected. - ``MetadataStore.load`` now reads chunks through a ``_chunk_meta_from_payload`` helper that drops unknown payload keys, keeping msgpack records both forward- and backward-compatible as later production stages add more fields. - ``MetadataStore.touch`` updates the stats and ``ServerCore.lookup`` calls it for every matched chunk, so hit traffic alone is enough to drive eviction and metrics later. - Unit tests cover the new defaults, the touch contract, both msgpack compatibility directions, and the existing lookup paths.
Align Stage A lookup tests with master prefix index helpers after rebase.
Expose opt-in Prometheus metrics and diagnostic cache explanations for the HTTP service, while keeping diagnostic lookups read-only so access counters reflect production demand. Add a baseline measurement script and tests for metrics, diagnostics, CLI wiring, and lookup access accounting.
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.
Summary
/metrics,/diag/explain, and a baseline measurement script.--enable-metricsintopython -m daser.serverand add tests for metrics, diagnostics, CLI behavior, msgpack compatibility, and lookup access accounting.Test plan
ruff check daser/ tests/ruff format --check daser/ tests/ruff check examples/baseline_measure/measure.pyruff format --check examples/baseline_measure/measure.pyPYTHONHASHSEED=0 /data/sza/conda_env/daser/bin/python -m pytest -q -m "not integration" --ignore=tests/integration --ignore=tests/connector/test_daser_connector.py --ignore=tests/connector/test_gds_transfer.py tests/(215 passed, 1 skipped)/health,/documents,/diag/explain,/infer, and/metricspassing; store-dir/data/sza/daser-http-smoke-stageaNotes
--enable-metricsis passed./data/sza/vllm_daser_smoke_8011.logand/data/sza/daser_http_smoke_2036.log.