[ML] Use internal user for internal inference action (#128327)

* Using correct origin for inference action

* Update docs/changelog/128327.yaml

* [CI] Auto commit changes from spotless

---------

Co-authored-by: elasticsearchmachine <infra-root+elasticsearchmachine@elastic.co>
This commit is contained in:
Jonathan Buttner 2025-05-22 14:59:12 -04:00 committed by GitHub
parent a3bff18ec1
commit 19e18a94a6
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
3 changed files with 13 additions and 2 deletions

View File

@ -0,0 +1,5 @@
pr: 128327
summary: Use internal user for internal inference action
area: Machine Learning
type: bug
issues: []

View File

@ -21,6 +21,9 @@ import java.util.List;
import java.util.Set;
import java.util.stream.Collectors;
import static org.elasticsearch.xpack.core.ClientHelper.INFERENCE_ORIGIN;
import static org.elasticsearch.xpack.core.ClientHelper.executeAsyncWithOrigin;
public class InferenceRunner {
private final Client client;
@ -73,6 +76,6 @@ public class InferenceRunner {
}
public void doInference(InferenceAction.Request request, ActionListener<InferenceAction.Response> listener) {
client.execute(InferenceAction.INSTANCE, request, listener);
executeAsyncWithOrigin(client, INFERENCE_ORIGIN, InferenceAction.INSTANCE, request, listener);
}
}

View File

@ -25,6 +25,9 @@ import java.util.Arrays;
import java.util.List;
import java.util.Map;
import static org.elasticsearch.xpack.core.ClientHelper.INFERENCE_ORIGIN;
import static org.elasticsearch.xpack.core.ClientHelper.executeAsyncWithOrigin;
/**
* A {@code RankFeaturePhaseRankCoordinatorContext} that performs a rerank inference call to determine relevance scores for documents within
* the provided rank window.
@ -114,7 +117,7 @@ public class TextSimilarityRankFeaturePhaseRankCoordinatorContext extends RankFe
List<String> featureData = Arrays.stream(featureDocs).map(x -> x.featureData).toList();
InferenceAction.Request inferenceRequest = generateRequest(featureData);
try {
client.execute(InferenceAction.INSTANCE, inferenceRequest, inferenceListener);
executeAsyncWithOrigin(client, INFERENCE_ORIGIN, InferenceAction.INSTANCE, inferenceRequest, inferenceListener);
} finally {
inferenceRequest.decRef();
}