The kNN visitedLimit javadoc (LeafReader#searchNearestVectors, TopKnnCollector, KnnCollectorManager, TopKnnCollectorManager, HnswGraphSearcher) calls it 'the maximum number of nodes that the search is allowed to visit', but since 10.3/10.4 batched scoring deliberately treats it as a best-effort limit that a search can exceed (javadoc only; the code behaviour is intended upstream)

Finding LFV-149Area: search / kNN visit limit (visitedLimit javadoc on LeafReader, TopKnnCollector and the kNN collector managers; HNSW and exhaustive vector search)Priority: lowImpact: minorNovelty: known upstream

AI-assisted finding, confirmed by automated checks, not yet triaged by a human Lucene committer. This site files nothing upstream.

What it shows

You call LeafReader#searchNearestVectors (or KnnVectorsReader#search with a TopKnnCollector) with visitedLimit L and expect what the javadoc says: L is 'the maximum number of nodes that the search is allowed to visit', so at most L vectors are scored. On 10.4.0 and later the search can score more: in the repro, a 50-vector default-codec segment with L = 1 scores all 50 vectors, an HNSW segment with L = 8 scores 20, and a Seeded search with 100 seeds and L = 2 scores 100, because vectors and entry points are scored in batches and counted only afterwards. Upstream made the limit best effort on purpose (#15057, #15075, #15500) and no wrong hits come back; only the public javadoc was never updated.

Expected vs actual

Expected: The test passes: it asserts the documented behaviour.
Actual: While the reported behaviour is present the test fails; the failure message states expected against actual.

Files

src/org/apache/lucene/search/TestExhaustiveKnnVisitedLimit.java (JUnit test)
src/org/apache/lucene/search/TestHnswEntryPhaseVisitedLimit.java (JUnit test)

Run it

Tested on Lucene main e1faeba08bde (2026-11-02): still reproduces.
Tested on Lucene main ea57360912ed (2026-10-24): reproduced.
How to read the outcome: a failing test is the finding reproducing (the failure message states expected against actual).
Seeds: a rerun with the same seed replays the same random choices. Assertions off (-da -dsa -Dtests.asserts=false) is the production-like setting.

# Lucene checkout at the tested commit, built with: ./gradlew --no-daemon assemble   (JDK 25)
LUCENE=/path/to/lucene; V=12.0.0-SNAPSHOT
CP=$LUCENE/lucene/core/build/libs/lucene-core-$V.jar
CP=$CP:$LUCENE/lucene/codecs/build/libs/lucene-codecs-$V.jar
CP=$CP:$LUCENE/lucene/test-framework/build/libs/lucene-test-framework-$V.jar
CP=$CP:randomizedtesting-runner-2.9.2.jar:junit-4.13.2.jar:hamcrest-3.0.jar   # test-framework dependencies (Maven Central)
CP=$CP:junit-platform-console-standalone-6.1.3.jar   # JUnit Platform console launcher (Maven Central)

javac -d classes -cp "$CP" $(find src -name '*.java')

# run 1: assertions on, seed DEADBEEF
java -ea -esa --add-modules jdk.incubator.vector --enable-native-access=ALL-UNNAMED -Dtests.seed=DEADBEEF \
     -cp "classes:$CP" org.junit.platform.console.ConsoleLauncher \
     execute --select-class org.apache.lucene.search.TestExhaustiveKnnVisitedLimit \
       --select-class org.apache.lucene.search.TestHnswEntryPhaseVisitedLimit

Gradle alternative: copy the test into the matching module's src/test/ tree of the checkout and run ./gradlew -p lucene/<module> test --tests <class> -Ptests.seed=<seed> (add -Ptests.asserts=false for a run without assertions).

Feedback

Open an issue in lucene-ai-bug-hunt. Please do not open issues or pull requests in apache/lucene about this finding unless a Lucene committer agrees it should be filed.

Snapshot of 2026-11-04. AI-assisted; not yet triaged by a human Lucene committer. Feedback: open an issue in lucene-ai-bug-hunt.