[8.x] Add node feature for failure store, refactor capability names (#126885) (#127091)

* Add node feature for failure store, refactor capability names (#126885)

Adds a node feature that is conditionally added to the cluster state if the failure store
feature flag is enabled. Requires all nodes in the cluster to have the node feature
present in order to redirect failed documents to the failure store from the ingest node
or from shard level bulk failures.

* Fix backporting issues

Author: jbaiera

modules/data-streams/src/main/java/org/elasticsearch/datastreams/DataStreamFeatures.java

@@ -12,6 +12,7 @@
 import org.elasticsearch.Version;
 import org.elasticsearch.action.admin.indices.rollover.LazyRolloverAction;
 import org.elasticsearch.action.datastreams.autosharding.DataStreamAutoShardingService;
+import org.elasticsearch.cluster.metadata.DataStream;
 import org.elasticsearch.cluster.metadata.DataStreamGlobalRetention;
 import org.elasticsearch.datastreams.lifecycle.health.DataStreamLifecycleHealthInfoPublisher;
 import org.elasticsearch.features.FeatureSpecification;
@@ -39,12 +40,22 @@ public Map<NodeFeature, Version> getHistoricalFeatures() {
 
     @Override
     public Set<NodeFeature> getFeatures() {
-        return Set.of(
-            DataStreamLifecycleHealthInfoPublisher.DSL_HEALTH_INFO_FEATURE,  // Added in 8.12
-            LazyRolloverAction.DATA_STREAM_LAZY_ROLLOVER,                    // Added in 8.13
-            DataStreamAutoShardingService.DATA_STREAM_AUTO_SHARDING_FEATURE,
-            DataStreamGlobalRetention.GLOBAL_RETENTION                       // Added in 8.14
-        );
+        if (DataStream.isFailureStoreFeatureFlagEnabled()) {
+            return Set.of(
+                DataStreamLifecycleHealthInfoPublisher.DSL_HEALTH_INFO_FEATURE,  // Added in 8.12
+                LazyRolloverAction.DATA_STREAM_LAZY_ROLLOVER,                    // Added in 8.13
+                DataStreamAutoShardingService.DATA_STREAM_AUTO_SHARDING_FEATURE,
+                DataStreamGlobalRetention.GLOBAL_RETENTION,                      // Added in 8.14
+                DataStream.DATA_STREAM_FAILURE_STORE_FEATURE                     // Added in 8.19
+            );
+        } else {
+            return Set.of(
+                DataStreamLifecycleHealthInfoPublisher.DSL_HEALTH_INFO_FEATURE,  // Added in 8.12
+                LazyRolloverAction.DATA_STREAM_LAZY_ROLLOVER,                    // Added in 8.13
+                DataStreamAutoShardingService.DATA_STREAM_AUTO_SHARDING_FEATURE,
+                DataStreamGlobalRetention.GLOBAL_RETENTION                       // Added in 8.14
+            );
+        }
     }
 
     @Override

modules/data-streams/src/yamlRestTest/resources/rest-api-spec/test/data_stream/10_basic.yml

@@ -217,7 +217,7 @@ setup:
       capabilities:
         - method: POST
           path: /_index_template/{template}
-          capabilities: [ 'failure_store_in_template' ]
+          capabilities: [ 'data_stream_options.failure_store' ]
 
   - do:
       ingest.put_pipeline:
@@ -643,7 +643,7 @@ setup:
       capabilities:
         - method: POST
           path: /_index_template/{template}
-          capabilities: [ 'failure_store_in_template' ]
+          capabilities: [ 'data_stream_options.failure_store' ]
 
   - do:
       allowed_warnings:
@@ -739,7 +739,7 @@ setup:
       capabilities:
         - method: POST
           path: /_index_template/{template}
-          capabilities: [ 'failure_store_in_template' ]
+          capabilities: [ 'data_stream_options.failure_store' ]
 
   - do:
       allowed_warnings:

modules/data-streams/src/yamlRestTest/resources/rest-api-spec/test/data_stream/150_tsdb.yml

@@ -190,7 +190,7 @@ TSDB failures go to failure store:
       capabilities:
         - method: POST
           path: /_index_template/{template}
-          capabilities: [ 'failure_store_in_template' ]
+          capabilities: [ 'data_stream_options.failure_store' ]
   - do:
       allowed_warnings:
         - "index template [my-template2] has index patterns [fs-k8s*] matching patterns from existing older templates [global] with patterns (global => [*]); this template [my-template2] will take precedence during new index creation"

modules/data-streams/src/yamlRestTest/resources/rest-api-spec/test/data_stream/170_modify_data_stream.yml

@@ -97,7 +97,7 @@
       capabilities:
         - method: POST
           path: /_index_template/{template}
-          capabilities: [ 'failure_store_in_template' ]
+          capabilities: [ 'data_stream_options.failure_store' ]
 
   - do:
       allowed_warnings:

modules/data-streams/src/yamlRestTest/resources/rest-api-spec/test/data_stream/190_failure_store_redirection.yml

@@ -8,7 +8,7 @@ setup:
           capabilities: [ 'failure_store_status' ]
         - method: POST
           path: /_index_template/{template}
-          capabilities: [ 'failure_store_in_template' ]
+          capabilities: [ 'data_stream_options.failure_store' ]
 
 ---
 teardown:

modules/data-streams/src/yamlRestTest/resources/rest-api-spec/test/data_stream/210_rollover_failure_store.yml

@@ -6,10 +6,10 @@ setup:
       capabilities:
         - method: POST
           path: /_index_template/{template}
-          capabilities: [ 'failure_store_in_template' ]
+          capabilities: [ 'data_stream_options.failure_store' ]
         - method: POST
           path: /{index}/_rollover
-          capabilities: [ 'lazy-rollover-failure-store', 'index-expression-selectors' ]
+          capabilities: [ 'index_expression_selectors' ]
 
   - do:
       allowed_warnings:
@@ -313,7 +313,7 @@ teardown:
       capabilities:
         - method: POST
           path: /{index}/_rollover
-          capabilities: [lazy-rollover-failure-store]
+          capabilities: [index_expression_selectors]
 
   # Initialize failure store
   - do:

modules/data-streams/src/yamlRestTest/resources/rest-api-spec/test/data_stream/220_failure_store_cluster_setting.yml

@@ -1,17 +1,15 @@
 setup:
   - requires:
-      reason: "Data stream options was added in 8.18+"
+      cluster_features: [ "data_stream.failure_store" ]
+      reason: "Failure store GA in 8.19+"
       test_runner_features: [ capabilities, allowed_warnings, contains ]
       capabilities:
         - method: POST
           path: /{index}/_doc
           capabilities: [ 'failure_store_status' ]
         - method: POST
           path: /_index_template/{template}
-          capabilities: [ 'failure_store_in_template' ]
-        - method: PUT
-          path: /_cluster/settings
-          capabilities: [ 'data_stream_failure_store_cluster_setting' ]
+          capabilities: [ 'data_stream_options.failure_store' ]
 
   - do:
       cluster.put_settings:

modules/data-streams/src/yamlRestTest/resources/rest-api-spec/test/data_stream/230_data_stream_options.yml

@@ -6,7 +6,7 @@ setup:
       capabilities:
         - method: POST
           path: /_index_template/{template}
-          capabilities: [ 'failure_store_in_template' ]
+          capabilities: [ 'data_stream_options.failure_store' ]
 
   - do:
       allowed_warnings:

modules/data-streams/src/yamlRestTest/resources/rest-api-spec/test/data_stream/30_auto_create_data_stream.yml

@@ -55,7 +55,7 @@
       capabilities:
         - method: POST
           path: /_index_template/{template}
-          capabilities: [ 'failure_store_in_template' ]
+          capabilities: [ 'data_stream_options.failure_store' ]
 
   - do:
       allowed_warnings:

server/src/main/java/org/elasticsearch/action/bulk/BulkOperation.java

@@ -99,6 +99,7 @@ final class BulkOperation extends ActionRunnable<BulkResponse> {
     private final Map<ShardId, Exception> shortCircuitShardFailures = ConcurrentCollections.newConcurrentMap();
     private final FailureStoreMetrics failureStoreMetrics;
     private final DataStreamFailureStoreSettings dataStreamFailureStoreSettings;
+    private final boolean clusterHasFailureStoreFeature;
 
     BulkOperation(
         Task task,
@@ -113,7 +114,8 @@ final class BulkOperation extends ActionRunnable<BulkResponse> {
         long startTimeNanos,
         ActionListener<BulkResponse> listener,
         FailureStoreMetrics failureStoreMetrics,
-        DataStreamFailureStoreSettings dataStreamFailureStoreSettings
+        DataStreamFailureStoreSettings dataStreamFailureStoreSettings,
+        boolean clusterHasFailureStoreFeature
     ) {
         this(
             task,
@@ -130,7 +132,8 @@ final class BulkOperation extends ActionRunnable<BulkResponse> {
             new ClusterStateObserver(clusterService, bulkRequest.timeout(), logger, threadPool.getThreadContext()),
             new FailureStoreDocumentConverter(),
             failureStoreMetrics,
-            dataStreamFailureStoreSettings
+            dataStreamFailureStoreSettings,
+            clusterHasFailureStoreFeature
         );
     }
 
@@ -149,7 +152,8 @@ final class BulkOperation extends ActionRunnable<BulkResponse> {
         ClusterStateObserver observer,
         FailureStoreDocumentConverter failureStoreDocumentConverter,
         FailureStoreMetrics failureStoreMetrics,
-        DataStreamFailureStoreSettings dataStreamFailureStoreSettings
+        DataStreamFailureStoreSettings dataStreamFailureStoreSettings,
+        boolean clusterHasFailureStoreFeature
     ) {
         super(listener);
         this.task = task;
@@ -169,6 +173,7 @@ final class BulkOperation extends ActionRunnable<BulkResponse> {
         this.shortCircuitShardFailures.putAll(bulkRequest.incrementalState().shardLevelFailures());
         this.failureStoreMetrics = failureStoreMetrics;
         this.dataStreamFailureStoreSettings = dataStreamFailureStoreSettings;
+        this.clusterHasFailureStoreFeature = clusterHasFailureStoreFeature;
     }
 
     @Override
@@ -543,7 +548,7 @@ private IndexDocFailureStoreStatus processFailure(BulkItemRequest bulkItemReques
         DataStream failureStoreCandidate = getRedirectTargetCandidate(docWriteRequest, clusterState.metadata());
         // If the candidate is not null, the BulkItemRequest targets a data stream, but we'll still have to check if
         // it has the failure store enabled.
-        if (failureStoreCandidate != null) {
+        if (failureStoreCandidate != null && clusterHasFailureStoreFeature) {
             // Do not redirect documents to a failure store that were already headed to one.
             var isFailureStoreRequest = isFailureStoreRequest(docWriteRequest);
             if (isFailureStoreRequest == false