[GitHub] [druid] clintropolis commented on a change in pull request #11006: [BUG FIX]Kinesis lag keep increasing when there is no more new data for kinesis stream

[GitBox <gi...@apache.org>]

clintropolis commented on a change in pull request #11006:
URL: https://github.com/apache/druid/pull/11006#discussion_r595748239



##########
File path: extensions-core/kinesis-indexing-service/src/main/java/org/apache/druid/indexing/kinesis/KinesisRecordSupplier.java
##########
@@ -863,21 +863,36 @@ private Long getPartitionTimeLag(StreamPartition<String> partition, String offse
         iteratorType = ShardIteratorType.AT_SEQUENCE_NUMBER.toString();
         offsetToUse = offset;
       }
-      String shardIterator = kinesis.getShardIterator(
-          partition.getStream(),
-          partition.getPartitionId(),
-          iteratorType,
-          offsetToUse
-      ).getShardIterator();
 
-      GetRecordsResult recordsResult = kinesis.getRecords(
-          new GetRecordsRequest().withShardIterator(shardIterator).withLimit(1)
-      );
+      GetRecordsResult recordsResult = getRecords(ShardIteratorType.AFTER_SEQUENCE_NUMBER.toString(), offsetToUse, partition);
+
+      // If no more new data after offsetToUse, it means there is no lag for now.
+      // So report lag points as 0L.
+      if (recordsResult.getRecords().size() == 0) {
+        return 0L;
+      } else {
+        recordsResult = getRecords(iteratorType, offsetToUse, partition);
+      }
 
       return recordsResult.getMillisBehindLatest();
     });
   }
 
+  private GetRecordsResult getRecords(String iteratorType, String offsetToUse, StreamPartition<String> partition)

Review comment:
       nit: might be nice to name this method `getRecordsForLag` or something similar to indicate its limited purpose (the `.withLimit(1)` makes it probably not really useful for actually getting records, nor does it need to be because other methods are handling that)




----------------------------------------------------------------
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.

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



---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@druid.apache.org
For additional commands, e-mail: commits-help@druid.apache.org