Examples with ADDataMigrator org.opensearch.ad.cluster.ADDataMigrator used on opensource projects

Example 1 with ADDataMigrator

use of org.opensearch.ad.cluster.ADDataMigrator in project anomaly-detection by opensearch-project.

the class AnomalyDetectorPlugin method createComponents.

@Override
public Collection<Object> createComponents(Client client, ClusterService clusterService, ThreadPool threadPool, ResourceWatcherService resourceWatcherService, ScriptService scriptService, NamedXContentRegistry xContentRegistry, Environment environment, NodeEnvironment nodeEnvironment, NamedWriteableRegistry namedWriteableRegistry, IndexNameExpressionResolver indexNameExpressionResolver, Supplier<RepositoriesService> repositoriesServiceSupplier) {
    EnabledSetting.getInstance().init(clusterService);
    NumericSetting.getInstance().init(clusterService);
    this.client = client;
    this.threadPool = threadPool;
    Settings settings = environment.settings();
    Throttler throttler = new Throttler(getClock());
    this.clientUtil = new ClientUtil(settings, client, throttler, threadPool);
    this.indexUtils = new IndexUtils(client, clientUtil, clusterService, indexNameExpressionResolver);
    this.nodeFilter = new DiscoveryNodeFilterer(clusterService);
    this.anomalyDetectionIndices = new AnomalyDetectionIndices(client, clusterService, threadPool, settings, nodeFilter, AnomalyDetectorSettings.MAX_UPDATE_RETRY_TIMES);
    this.clusterService = clusterService;
    SingleFeatureLinearUniformInterpolator singleFeatureLinearUniformInterpolator = new IntegerSensitiveSingleFeatureLinearUniformInterpolator();
    Interpolator interpolator = new LinearUniformInterpolator(singleFeatureLinearUniformInterpolator);
    SearchFeatureDao searchFeatureDao = new SearchFeatureDao(client, xContentRegistry, interpolator, clientUtil, settings, clusterService, AnomalyDetectorSettings.NUM_SAMPLES_PER_TREE);
    JvmService jvmService = new JvmService(environment.settings());
    RandomCutForestMapper mapper = new RandomCutForestMapper();
    mapper.setSaveExecutorContextEnabled(true);
    mapper.setSaveTreeStateEnabled(true);
    mapper.setPartialTreeStateEnabled(true);
    V1JsonToV2StateConverter converter = new V1JsonToV2StateConverter();
    double modelMaxSizePercent = AnomalyDetectorSettings.MODEL_MAX_SIZE_PERCENTAGE.get(settings);
    ADCircuitBreakerService adCircuitBreakerService = new ADCircuitBreakerService(jvmService).init();
    MemoryTracker memoryTracker = new MemoryTracker(jvmService, modelMaxSizePercent, AnomalyDetectorSettings.DESIRED_MODEL_SIZE_PERCENTAGE, clusterService, adCircuitBreakerService);
    NodeStateManager stateManager = new NodeStateManager(client, xContentRegistry, settings, clientUtil, getClock(), AnomalyDetectorSettings.HOURLY_MAINTENANCE, clusterService);
    FeatureManager featureManager = new FeatureManager(searchFeatureDao, interpolator, getClock(), AnomalyDetectorSettings.MAX_TRAIN_SAMPLE, AnomalyDetectorSettings.MAX_SAMPLE_STRIDE, AnomalyDetectorSettings.TRAIN_SAMPLE_TIME_RANGE_IN_HOURS, AnomalyDetectorSettings.MIN_TRAIN_SAMPLES, AnomalyDetectorSettings.MAX_SHINGLE_PROPORTION_MISSING, AnomalyDetectorSettings.MAX_IMPUTATION_NEIGHBOR_DISTANCE, AnomalyDetectorSettings.PREVIEW_SAMPLE_RATE, AnomalyDetectorSettings.MAX_PREVIEW_SAMPLES, AnomalyDetectorSettings.HOURLY_MAINTENANCE, threadPool, AD_THREAD_POOL_NAME);
    long heapSizeBytes = JvmInfo.jvmInfo().getMem().getHeapMax().getBytes();
    serializeRCFBufferPool = AccessController.doPrivileged(new PrivilegedAction<GenericObjectPool<LinkedBuffer>>() {

        @Override
        public GenericObjectPool<LinkedBuffer> run() {
            return new GenericObjectPool<>(new BasePooledObjectFactory<LinkedBuffer>() {

                @Override
                public LinkedBuffer create() throws Exception {
                    return LinkedBuffer.allocate(AnomalyDetectorSettings.SERIALIZATION_BUFFER_BYTES);
                }

                @Override
                public PooledObject<LinkedBuffer> wrap(LinkedBuffer obj) {
                    return new DefaultPooledObject<>(obj);
                }
            });
        }
    });
    serializeRCFBufferPool.setMaxTotal(AnomalyDetectorSettings.MAX_TOTAL_RCF_SERIALIZATION_BUFFERS);
    serializeRCFBufferPool.setMaxIdle(AnomalyDetectorSettings.MAX_TOTAL_RCF_SERIALIZATION_BUFFERS);
    serializeRCFBufferPool.setMinIdle(0);
    serializeRCFBufferPool.setBlockWhenExhausted(false);
    serializeRCFBufferPool.setTimeBetweenEvictionRuns(AnomalyDetectorSettings.HOURLY_MAINTENANCE);
    CheckpointDao checkpoint = new CheckpointDao(client, clientUtil, CommonName.CHECKPOINT_INDEX_NAME, gson, mapper, converter, new ThresholdedRandomCutForestMapper(), AccessController.doPrivileged((PrivilegedAction<Schema<ThresholdedRandomCutForestState>>) () -> RuntimeSchema.getSchema(ThresholdedRandomCutForestState.class)), HybridThresholdingModel.class, anomalyDetectionIndices, AnomalyDetectorSettings.MAX_CHECKPOINT_BYTES, serializeRCFBufferPool, AnomalyDetectorSettings.SERIALIZATION_BUFFER_BYTES, 1 - AnomalyDetectorSettings.THRESHOLD_MIN_PVALUE);
    Random random = new Random(42);
    CheckpointWriteWorker checkpointWriteQueue = new CheckpointWriteWorker(heapSizeBytes, AnomalyDetectorSettings.CHECKPOINT_WRITE_QUEUE_SIZE_IN_BYTES, AnomalyDetectorSettings.CHECKPOINT_WRITE_QUEUE_MAX_HEAP_PERCENT, clusterService, random, adCircuitBreakerService, threadPool, settings, AnomalyDetectorSettings.MAX_QUEUED_TASKS_RATIO, getClock(), AnomalyDetectorSettings.MEDIUM_SEGMENT_PRUNE_RATIO, AnomalyDetectorSettings.LOW_SEGMENT_PRUNE_RATIO, AnomalyDetectorSettings.MAINTENANCE_FREQ_CONSTANT, AnomalyDetectorSettings.QUEUE_MAINTENANCE, checkpoint, CommonName.CHECKPOINT_INDEX_NAME, AnomalyDetectorSettings.HOURLY_MAINTENANCE, stateManager, AnomalyDetectorSettings.HOURLY_MAINTENANCE);
    EntityCache cache = new PriorityCache(checkpoint, AnomalyDetectorSettings.DEDICATED_CACHE_SIZE.get(settings), AnomalyDetectorSettings.CHECKPOINT_TTL, AnomalyDetectorSettings.MAX_INACTIVE_ENTITIES, memoryTracker, AnomalyDetectorSettings.NUM_TREES, getClock(), clusterService, AnomalyDetectorSettings.HOURLY_MAINTENANCE, threadPool, checkpointWriteQueue, AnomalyDetectorSettings.MAINTENANCE_FREQ_CONSTANT);
    CacheProvider cacheProvider = new CacheProvider(cache);
    EntityColdStarter entityColdStarter = new EntityColdStarter(getClock(), threadPool, stateManager, AnomalyDetectorSettings.NUM_SAMPLES_PER_TREE, AnomalyDetectorSettings.NUM_TREES, AnomalyDetectorSettings.TIME_DECAY, AnomalyDetectorSettings.NUM_MIN_SAMPLES, AnomalyDetectorSettings.MAX_SAMPLE_STRIDE, AnomalyDetectorSettings.MAX_TRAIN_SAMPLE, interpolator, searchFeatureDao, AnomalyDetectorSettings.THRESHOLD_MIN_PVALUE, featureManager, settings, AnomalyDetectorSettings.HOURLY_MAINTENANCE, checkpointWriteQueue, AnomalyDetectorSettings.MAX_COLD_START_ROUNDS);
    EntityColdStartWorker coldstartQueue = new EntityColdStartWorker(heapSizeBytes, AnomalyDetectorSettings.ENTITY_REQUEST_SIZE_IN_BYTES, AnomalyDetectorSettings.ENTITY_COLD_START_QUEUE_MAX_HEAP_PERCENT, clusterService, random, adCircuitBreakerService, threadPool, settings, AnomalyDetectorSettings.MAX_QUEUED_TASKS_RATIO, getClock(), AnomalyDetectorSettings.MEDIUM_SEGMENT_PRUNE_RATIO, AnomalyDetectorSettings.LOW_SEGMENT_PRUNE_RATIO, AnomalyDetectorSettings.MAINTENANCE_FREQ_CONSTANT, AnomalyDetectorSettings.QUEUE_MAINTENANCE, entityColdStarter, AnomalyDetectorSettings.HOURLY_MAINTENANCE, stateManager);
    ModelManager modelManager = new ModelManager(checkpoint, getClock(), AnomalyDetectorSettings.NUM_TREES, AnomalyDetectorSettings.NUM_SAMPLES_PER_TREE, AnomalyDetectorSettings.TIME_DECAY, AnomalyDetectorSettings.NUM_MIN_SAMPLES, AnomalyDetectorSettings.THRESHOLD_MIN_PVALUE, AnomalyDetectorSettings.MIN_PREVIEW_SIZE, AnomalyDetectorSettings.HOURLY_MAINTENANCE, AnomalyDetectorSettings.HOURLY_MAINTENANCE, entityColdStarter, featureManager, memoryTracker);
    MultiEntityResultHandler multiEntityResultHandler = new MultiEntityResultHandler(client, settings, threadPool, anomalyDetectionIndices, this.clientUtil, this.indexUtils, clusterService);
    ResultWriteWorker resultWriteQueue = new ResultWriteWorker(heapSizeBytes, AnomalyDetectorSettings.RESULT_WRITE_QUEUE_SIZE_IN_BYTES, AnomalyDetectorSettings.RESULT_WRITE_QUEUE_MAX_HEAP_PERCENT, clusterService, random, adCircuitBreakerService, threadPool, settings, AnomalyDetectorSettings.MAX_QUEUED_TASKS_RATIO, getClock(), AnomalyDetectorSettings.MEDIUM_SEGMENT_PRUNE_RATIO, AnomalyDetectorSettings.LOW_SEGMENT_PRUNE_RATIO, AnomalyDetectorSettings.MAINTENANCE_FREQ_CONSTANT, AnomalyDetectorSettings.QUEUE_MAINTENANCE, multiEntityResultHandler, xContentRegistry, stateManager, AnomalyDetectorSettings.HOURLY_MAINTENANCE);
    CheckpointReadWorker checkpointReadQueue = new CheckpointReadWorker(heapSizeBytes, AnomalyDetectorSettings.ENTITY_FEATURE_REQUEST_SIZE_IN_BYTES, AnomalyDetectorSettings.CHECKPOINT_READ_QUEUE_MAX_HEAP_PERCENT, clusterService, random, adCircuitBreakerService, threadPool, settings, AnomalyDetectorSettings.MAX_QUEUED_TASKS_RATIO, getClock(), AnomalyDetectorSettings.MEDIUM_SEGMENT_PRUNE_RATIO, AnomalyDetectorSettings.LOW_SEGMENT_PRUNE_RATIO, AnomalyDetectorSettings.MAINTENANCE_FREQ_CONSTANT, AnomalyDetectorSettings.QUEUE_MAINTENANCE, modelManager, checkpoint, coldstartQueue, resultWriteQueue, stateManager, anomalyDetectionIndices, cacheProvider, AnomalyDetectorSettings.HOURLY_MAINTENANCE, checkpointWriteQueue);
    ColdEntityWorker coldEntityQueue = new ColdEntityWorker(heapSizeBytes, AnomalyDetectorSettings.ENTITY_FEATURE_REQUEST_SIZE_IN_BYTES, AnomalyDetectorSettings.COLD_ENTITY_QUEUE_MAX_HEAP_PERCENT, clusterService, random, adCircuitBreakerService, threadPool, settings, AnomalyDetectorSettings.MAX_QUEUED_TASKS_RATIO, getClock(), AnomalyDetectorSettings.MEDIUM_SEGMENT_PRUNE_RATIO, AnomalyDetectorSettings.LOW_SEGMENT_PRUNE_RATIO, AnomalyDetectorSettings.MAINTENANCE_FREQ_CONSTANT, checkpointReadQueue, AnomalyDetectorSettings.HOURLY_MAINTENANCE, stateManager);
    ADDataMigrator dataMigrator = new ADDataMigrator(client, clusterService, xContentRegistry, anomalyDetectionIndices);
    HashRing hashRing = new HashRing(nodeFilter, getClock(), settings, client, clusterService, dataMigrator, modelManager);
    anomalyDetectorRunner = new AnomalyDetectorRunner(modelManager, featureManager, AnomalyDetectorSettings.MAX_PREVIEW_RESULTS);
    Map<String, ADStat<?>> stats = ImmutableMap.<String, ADStat<?>>builder().put(StatNames.AD_EXECUTE_REQUEST_COUNT.getName(), new ADStat<>(false, new CounterSupplier())).put(StatNames.AD_EXECUTE_FAIL_COUNT.getName(), new ADStat<>(false, new CounterSupplier())).put(StatNames.AD_HC_EXECUTE_REQUEST_COUNT.getName(), new ADStat<>(false, new CounterSupplier())).put(StatNames.AD_HC_EXECUTE_FAIL_COUNT.getName(), new ADStat<>(false, new CounterSupplier())).put(StatNames.MODEL_INFORMATION.getName(), new ADStat<>(false, new ModelsOnNodeSupplier(modelManager, cacheProvider, settings, clusterService))).put(StatNames.ANOMALY_DETECTORS_INDEX_STATUS.getName(), new ADStat<>(true, new IndexStatusSupplier(indexUtils, AnomalyDetector.ANOMALY_DETECTORS_INDEX))).put(StatNames.ANOMALY_RESULTS_INDEX_STATUS.getName(), new ADStat<>(true, new IndexStatusSupplier(indexUtils, CommonName.ANOMALY_RESULT_INDEX_ALIAS))).put(StatNames.MODELS_CHECKPOINT_INDEX_STATUS.getName(), new ADStat<>(true, new IndexStatusSupplier(indexUtils, CommonName.CHECKPOINT_INDEX_NAME))).put(StatNames.ANOMALY_DETECTION_JOB_INDEX_STATUS.getName(), new ADStat<>(true, new IndexStatusSupplier(indexUtils, AnomalyDetectorJob.ANOMALY_DETECTOR_JOB_INDEX))).put(StatNames.ANOMALY_DETECTION_STATE_STATUS.getName(), new ADStat<>(true, new IndexStatusSupplier(indexUtils, CommonName.DETECTION_STATE_INDEX))).put(StatNames.DETECTOR_COUNT.getName(), new ADStat<>(true, new SettableSupplier())).put(StatNames.SINGLE_ENTITY_DETECTOR_COUNT.getName(), new ADStat<>(true, new SettableSupplier())).put(StatNames.MULTI_ENTITY_DETECTOR_COUNT.getName(), new ADStat<>(true, new SettableSupplier())).put(StatNames.AD_EXECUTING_BATCH_TASK_COUNT.getName(), new ADStat<>(false, new CounterSupplier())).put(StatNames.AD_CANCELED_BATCH_TASK_COUNT.getName(), new ADStat<>(false, new CounterSupplier())).put(StatNames.AD_TOTAL_BATCH_TASK_EXECUTION_COUNT.getName(), new ADStat<>(false, new CounterSupplier())).put(StatNames.AD_BATCH_TASK_FAILURE_COUNT.getName(), new ADStat<>(false, new CounterSupplier())).put(StatNames.MODEL_COUNT.getName(), new ADStat<>(false, new ModelsOnNodeCountSupplier(modelManager, cacheProvider))).build();
    adStats = new ADStats(stats);
    adTaskCacheManager = new ADTaskCacheManager(settings, clusterService, memoryTracker);
    adTaskManager = new ADTaskManager(settings, clusterService, client, xContentRegistry, anomalyDetectionIndices, nodeFilter, hashRing, adTaskCacheManager, threadPool);
    AnomalyResultBulkIndexHandler anomalyResultBulkIndexHandler = new AnomalyResultBulkIndexHandler(client, settings, threadPool, this.clientUtil, this.indexUtils, clusterService, anomalyDetectionIndices);
    adBatchTaskRunner = new ADBatchTaskRunner(settings, threadPool, clusterService, client, adCircuitBreakerService, featureManager, adTaskManager, anomalyDetectionIndices, adStats, anomalyResultBulkIndexHandler, adTaskCacheManager, searchFeatureDao, hashRing, modelManager);
    ADSearchHandler adSearchHandler = new ADSearchHandler(settings, clusterService, client);
    // transport action handler constructors
    return ImmutableList.of(anomalyDetectionIndices, anomalyDetectorRunner, searchFeatureDao, singleFeatureLinearUniformInterpolator, interpolator, gson, jvmService, hashRing, featureManager, modelManager, stateManager, new ADClusterEventListener(clusterService, hashRing), adCircuitBreakerService, adStats, new MasterEventListener(clusterService, threadPool, client, getClock(), clientUtil, nodeFilter), nodeFilter, multiEntityResultHandler, checkpoint, cacheProvider, adTaskManager, adBatchTaskRunner, adSearchHandler, coldstartQueue, resultWriteQueue, checkpointReadQueue, checkpointWriteQueue, coldEntityQueue, entityColdStarter, adTaskCacheManager);
}