jugomezv commented on code in PR #9994:
URL: https://github.com/apache/pinot/pull/9994#discussion_r1060944878


##########
pinot-core/src/main/java/org/apache/pinot/core/data/manager/realtime/ConsumptionDelayTracker.java:
##########
@@ -0,0 +1,361 @@
+/**
+ * Licensed to the Apache Software Foundation (ASF) under one
+ * or more contributor license agreements.  See the NOTICE file
+ * distributed with this work for additional information
+ * regarding copyright ownership.  The ASF licenses this file
+ * to you under the Apache License, Version 2.0 (the
+ * "License"); you may not use this file except in compliance
+ * with the License.  You may obtain a copy of the License at
+ *
+ *   http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing,
+ * software distributed under the License is distributed on an
+ * "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+ * KIND, either express or implied.  See the License for the
+ * specific language governing permissions and limitations
+ * under the License.
+ */
+
+package org.apache.pinot.core.data.manager.realtime;
+
+import com.google.common.annotations.VisibleForTesting;
+import java.util.HashSet;
+import java.util.List;
+import java.util.Timer;
+import java.util.TimerTask;
+import java.util.concurrent.ConcurrentHashMap;
+import org.apache.pinot.common.metrics.ServerGauge;
+import org.apache.pinot.common.metrics.ServerMetrics;
+import org.slf4j.Logger;
+import org.slf4j.LoggerFactory;
+
+
+/**
+ * A Class to track maximum realtime delay for a given table on a given server.
+ * Highlights:
+ * 1-An object of this class is hosted by each RealtimeTableDataManager.
+ * 2-The object tracks ingestion delays for all partitions hosted by the 
current server for the given Realtime table.
+ * 3-Partition delays are updated by all LLRealtimeSegmentDataManager objects 
hosted in the corresponding
+ *   RealtimeTableDataManager.
+ * 4-The class tracks the maximum of all ingestion delays observed for all 
partitions of the given table.
+ * 5-A Metric is derived from reading the maximum tracked by this class.
+ * 6-Delay reported for partitions that do not have events to consume is 
reported as zero.
+ * 7-We track the time at which each delay sample was collected so that delay 
can be increased when partition stops
+ *   consuming for any reason other than no events being available for 
consumption.
+ * 8-Segments that go from CONSUMING to DROPPED states stop being tracked so 
their delays do not cloud
+ *   delays of active partitions.
+ * 9-When a segment goes from CONSUMING to ONLINE, we start ticking time for 
the corresponding partition.
+ *   If no consumption is noticed after a timeout, then we read ideal state to 
confirm the server still hosts the
+ *   partition. If not, we stop tracking the respective partition.
+ * 10-A timer thread is started by this object to track timeouts of partitions 
and drive the reading of their ideal
+ *    state.
+ *
+ *  The following diagram illustrates the object interactions with main 
external APIs
+ *
+ *     (CONSUMING -> ONLINE state change)
+ *             |
+ *      markPartitionForConfirmation(partitionId)
+ *            |                         
|<-storeConsumptionDelay()-{LLRealtimeSegmentDataManager(Partition 0}}
+ *            |                         |
+ * ___________V_________________________V_
+ * |           (Table X)                
|<-storeConsumptionDelay()-{LLRealtimeSegmentDataManager(Partition 1}}
+ * | ConsumptionDelayTracker            |           ...
+ * 
|____________________________________|<-storeConsumptionDelay()-{LLRealtimeSegmentDataManager
 (Partition n}}
+ *              ^                      ^
+ *              |                       \
+ *   timeoutInactivePartitions()    
stopTrackingPartitionConsumptionDelay(partitionId)
+ *    _________|__________                \
+ *   | TimerTrackingTask |          (CONSUMING -> DROPPED state change)
+ *   |___________________|
+ *
+ */
+
+public class ConsumptionDelayTracker {
+
+  // Sleep interval for timer thread that triggers read of ideal state
+  private static final int TIMER_THREAD_TICK_INTERVAL_MS = 300000; // 5 
minutes +/- precision in timeouts
+  // Once a partition is marked for verification, we wait 10 minutes to pull 
its ideal state.
+  private static final int PARTITION_TIMEOUT_MS = 600000;          // 10 
minutes timeouts
+  // Delay Timer thread for this time after starting timer
+  private static final int INITIAL_TIMER_THREAD_DELAY_MS = 100;

Review Comment:
   We need to pass a initial delay for the call.



-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: commits-unsubscr...@pinot.apache.org

For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscr...@pinot.apache.org
For additional commands, e-mail: commits-h...@pinot.apache.org

Reply via email to