elastic · jan-elastic · Sep 12, 2024 · Sep 12, 2024
diff --git a/test/test-clusters/src/main/java/org/elasticsearch/test/cluster/FeatureFlag.java b/test/test-clusters/src/main/java/org/elasticsearch/test/cluster/FeatureFlag.java
@@ -16,12 +16,7 @@
  */
 public enum FeatureFlag {
     TIME_SERIES_MODE("es.index_mode_feature_flag_registered=true", Version.fromString("8.0.0"), null),
-    FAILURE_STORE_ENABLED("es.failure_store_feature_flag_enabled=true", Version.fromString("8.12.0"), null),
-    INFERENCE_ADAPTIVE_ALLOCATIONS_ENABLED(
-        "es.inference_adaptive_allocations_feature_flag_enabled=true",
-        Version.fromString("8.16.0"),
-        null
-    );
+    FAILURE_STORE_ENABLED("es.failure_store_feature_flag_enabled=true", Version.fromString("8.12.0"), null);
 
     public final String systemProperty;
     public final Version from;

diff --git a/...c/main/java/org/elasticsearch/xpack/core/ml/action/StartTrainedModelDeploymentAction.java b/...c/main/java/org/elasticsearch/xpack/core/ml/action/StartTrainedModelDeploymentAction.java
@@ -29,7 +29,6 @@
 import org.elasticsearch.xcontent.XContentParser;
 import org.elasticsearch.xpack.core.ml.MlConfigVersion;
 import org.elasticsearch.xpack.core.ml.inference.TrainedModelConfig;
-import org.elasticsearch.xpack.core.ml.inference.assignment.AdaptiveAllocationsFeatureFlag;
 import org.elasticsearch.xpack.core.ml.inference.assignment.AdaptiveAllocationsSettings;
 import org.elasticsearch.xpack.core.ml.inference.assignment.AllocationStatus;
 import org.elasticsearch.xpack.core.ml.inference.assignment.Priority;
@@ -120,14 +119,12 @@ public static class Request extends MasterNodeRequest<Request> implements ToXCon
                 ObjectParser.ValueType.VALUE
             );
             PARSER.declareString(Request::setPriority, PRIORITY);
-            if (AdaptiveAllocationsFeatureFlag.isEnabled()) {
-                PARSER.declareObjectOrNull(
-                    Request::setAdaptiveAllocationsSettings,
-                    (p, c) -> AdaptiveAllocationsSettings.PARSER.parse(p, c).build(),
-                    null,
-                    ADAPTIVE_ALLOCATIONS
-                );
-            }
+            PARSER.declareObjectOrNull(
+                Request::setAdaptiveAllocationsSettings,
+                (p, c) -> AdaptiveAllocationsSettings.PARSER.parse(p, c).build(),
+                null,
+                ADAPTIVE_ALLOCATIONS
+            );
         }
 
         public static Request parseRequest(String modelId, String deploymentId, XContentParser parser) {

diff --git a/.../main/java/org/elasticsearch/xpack/core/ml/action/UpdateTrainedModelDeploymentAction.java b/.../main/java/org/elasticsearch/xpack/core/ml/action/UpdateTrainedModelDeploymentAction.java
@@ -20,7 +20,6 @@
 import org.elasticsearch.xcontent.ToXContentObject;
 import org.elasticsearch.xcontent.XContentBuilder;
 import org.elasticsearch.xcontent.XContentParser;
-import org.elasticsearch.xpack.core.ml.inference.assignment.AdaptiveAllocationsFeatureFlag;
 import org.elasticsearch.xpack.core.ml.inference.assignment.AdaptiveAllocationsSettings;
 import org.elasticsearch.xpack.core.ml.job.messages.Messages;
 import org.elasticsearch.xpack.core.ml.utils.ExceptionsHelper;
@@ -50,14 +49,12 @@ public static class Request extends AcknowledgedRequest<Request> implements ToXC
         static {
             PARSER.declareString(Request::setDeploymentId, MODEL_ID);
             PARSER.declareInt(Request::setNumberOfAllocations, NUMBER_OF_ALLOCATIONS);
-            if (AdaptiveAllocationsFeatureFlag.isEnabled()) {
-                PARSER.declareObjectOrNull(
-                    Request::setAdaptiveAllocationsSettings,
-                    (p, c) -> AdaptiveAllocationsSettings.PARSER.parse(p, c).build(),
-                    AdaptiveAllocationsSettings.RESET_PLACEHOLDER,
-                    ADAPTIVE_ALLOCATIONS
-                );
-            }
+            PARSER.declareObjectOrNull(
+                Request::setAdaptiveAllocationsSettings,
+                (p, c) -> AdaptiveAllocationsSettings.PARSER.parse(p, c).build(),
+                AdaptiveAllocationsSettings.RESET_PLACEHOLDER,
+                ADAPTIVE_ALLOCATIONS
+            );
             PARSER.declareString((r, val) -> r.ackTimeout(TimeValue.parseTimeValue(val, TIMEOUT.getPreferredName())), TIMEOUT);
         }
 

diff --git a/.../org/elasticsearch/xpack/core/ml/inference/assignment/AdaptiveAllocationsFeatureFlag.java b/.../org/elasticsearch/xpack/core/ml/inference/assignment/AdaptiveAllocationsFeatureFlag.java
diff --git a/...ternalClusterTest/java/org/elasticsearch/xpack/inference/integration/ModelRegistryIT.java b/...ternalClusterTest/java/org/elasticsearch/xpack/inference/integration/ModelRegistryIT.java
@@ -25,7 +25,6 @@
 import org.elasticsearch.test.ESSingleNodeTestCase;
 import org.elasticsearch.xcontent.ToXContentObject;
 import org.elasticsearch.xcontent.XContentBuilder;
-import org.elasticsearch.xpack.core.ml.inference.assignment.AdaptiveAllocationsFeatureFlag;
 import org.elasticsearch.xpack.inference.InferencePlugin;
 import org.elasticsearch.xpack.inference.registry.ModelRegistry;
 import org.elasticsearch.xpack.inference.services.elser.ElserInternalModel;
@@ -102,7 +101,6 @@ public void testStoreModelWithUnknownFields() throws Exception {
     }
 
     public void testGetModel() throws Exception {
-        assumeTrue("Only if 'inference_adaptive_allocations' feature flag is enabled", AdaptiveAllocationsFeatureFlag.isEnabled());
         String inferenceEntityId = "test-get-model";
         Model model = buildElserModelConfig(inferenceEntityId, TaskType.SPARSE_EMBEDDING);
         AtomicReference<Boolean> putModelHolder = new AtomicReference<>();

diff --git a/...ugin/inference/src/main/java/org/elasticsearch/xpack/inference/services/ServiceUtils.java b/...ugin/inference/src/main/java/org/elasticsearch/xpack/inference/services/ServiceUtils.java
@@ -24,7 +24,6 @@
 import org.elasticsearch.xpack.core.inference.action.InferenceAction;
 import org.elasticsearch.xpack.core.inference.results.InferenceTextEmbeddingFloatResults;
 import org.elasticsearch.xpack.core.inference.results.TextEmbedding;
-import org.elasticsearch.xpack.core.ml.inference.assignment.AdaptiveAllocationsFeatureFlag;
 import org.elasticsearch.xpack.core.ml.inference.assignment.AdaptiveAllocationsSettings;
 import org.elasticsearch.xpack.inference.services.settings.ApiKeySecrets;
 
@@ -137,9 +136,6 @@ public static AdaptiveAllocationsSettings removeAsAdaptiveAllocationsSettings(
         String key,
         ValidationException validationException
     ) {
-        if (AdaptiveAllocationsFeatureFlag.isEnabled() == false) {
-            return null;
-        }
         Map<String, Object> settingsMap = ServiceUtils.removeFromMap(sourceMap, key);
         if (settingsMap == null) {
             return null;

diff --git a/...icsearch/xpack/inference/services/elasticsearch/ElasticsearchInternalServiceSettings.java b/...icsearch/xpack/inference/services/elasticsearch/ElasticsearchInternalServiceSettings.java
@@ -16,7 +16,6 @@
 import org.elasticsearch.inference.ServiceSettings;
 import org.elasticsearch.xcontent.ToXContentObject;
 import org.elasticsearch.xcontent.XContentBuilder;
-import org.elasticsearch.xpack.core.ml.inference.assignment.AdaptiveAllocationsFeatureFlag;
 import org.elasticsearch.xpack.core.ml.inference.assignment.AdaptiveAllocationsSettings;
 import org.elasticsearch.xpack.inference.services.ServiceUtils;
 
@@ -88,18 +87,12 @@ protected static ElasticsearchInternalServiceSettings.Builder fromMap(
         String modelId = extractOptionalString(map, MODEL_ID, ModelConfigurations.SERVICE_SETTINGS, validationException);
 
         if (numAllocations == null && adaptiveAllocationsSettings == null) {
-            if (AdaptiveAllocationsFeatureFlag.isEnabled()) {
-                validationException.addValidationError(
-                    ServiceUtils.missingOneOfSettingsErrorMsg(
-                        List.of(NUM_ALLOCATIONS, ADAPTIVE_ALLOCATIONS),
-                        ModelConfigurations.SERVICE_SETTINGS
-                    )
-                );
-            } else {
-                validationException.addValidationError(
-                    ServiceUtils.missingSettingErrorMsg(NUM_ALLOCATIONS, ModelConfigurations.SERVICE_SETTINGS)
-                );
-            }
+            validationException.addValidationError(
+                ServiceUtils.missingOneOfSettingsErrorMsg(
+                    List.of(NUM_ALLOCATIONS, ADAPTIVE_ALLOCATIONS),
+                    ModelConfigurations.SERVICE_SETTINGS
+                )
+            );
         }
 
         // if an error occurred while parsing, we'll set these to an invalid value, so we don't accidentally get a

diff --git a/...inference/src/test/java/org/elasticsearch/xpack/inference/services/ServiceUtilsTests.java b/...inference/src/test/java/org/elasticsearch/xpack/inference/services/ServiceUtilsTests.java
@@ -21,7 +21,6 @@
 import org.elasticsearch.test.ESTestCase;
 import org.elasticsearch.xpack.core.inference.results.InferenceTextEmbeddingByteResults;
 import org.elasticsearch.xpack.core.inference.results.InferenceTextEmbeddingFloatResults;
-import org.elasticsearch.xpack.core.ml.inference.assignment.AdaptiveAllocationsFeatureFlag;
 import org.elasticsearch.xpack.core.ml.inference.assignment.AdaptiveAllocationsSettings;
 import org.elasticsearch.xpack.inference.results.InferenceTextEmbeddingByteResultsTests;
 import org.elasticsearch.xpack.inference.results.TextEmbeddingResultsTests;
@@ -290,8 +289,6 @@ public void testRemoveAsOneOfTypesMissingReturnsNull() {
     }
 
     public void testRemoveAsAdaptiveAllocationsSettings() {
-        assumeTrue("Should only run if adaptive allocations feature flag is enabled", AdaptiveAllocationsFeatureFlag.isEnabled());
-
         Map<String, Object> map = new HashMap<>(
             Map.of("settings", new HashMap<>(Map.of("enabled", true, "min_number_of_allocations", 7, "max_number_of_allocations", 42)))
         );
@@ -314,8 +311,6 @@ public void testRemoveAsAdaptiveAllocationsSettings() {
     }
 
     public void testRemoveAsAdaptiveAllocationsSettings_exceptions() {
-        assumeTrue("Should only run if adaptive allocations feature flag is enabled", AdaptiveAllocationsFeatureFlag.isEnabled());
-
         Map<String, Object> map = new HashMap<>(
             Map.of("settings", new HashMap<>(Map.of("enabled", "YES!", "blah", 42, "max_number_of_allocations", -7)))
         );

diff --git a/x-pack/plugin/ml/src/main/java/org/elasticsearch/xpack/ml/MlInitializationService.java b/x-pack/plugin/ml/src/main/java/org/elasticsearch/xpack/ml/MlInitializationService.java
@@ -33,7 +33,6 @@
 import org.elasticsearch.telemetry.metric.MeterRegistry;
 import org.elasticsearch.threadpool.ThreadPool;
 import org.elasticsearch.xpack.core.ml.annotations.AnnotationIndex;
-import org.elasticsearch.xpack.core.ml.inference.assignment.AdaptiveAllocationsFeatureFlag;
 import org.elasticsearch.xpack.ml.inference.adaptiveallocations.AdaptiveAllocationsScalerService;
 import org.elasticsearch.xpack.ml.notifications.InferenceAuditor;
 
@@ -126,17 +125,13 @@ public void beforeStop() {
 
     public void onMaster() {
         mlDailyMaintenanceService.start();
-        if (AdaptiveAllocationsFeatureFlag.isEnabled()) {
-            adaptiveAllocationsScalerService.start();
-        }
+        adaptiveAllocationsScalerService.start();
         threadPool.executor(MachineLearning.UTILITY_THREAD_POOL_NAME).execute(this::makeMlInternalIndicesHidden);
     }
 
     public void offMaster() {
         mlDailyMaintenanceService.stop();
-        if (AdaptiveAllocationsFeatureFlag.isEnabled()) {
-            adaptiveAllocationsScalerService.stop();
-        }
+        adaptiveAllocationsScalerService.stop();
     }
 
     @Override