Fix NPE in segment replicator

IndexStatsIT.testConcurrentIndexingAndStatsRequests sometimes fails with
the following NPE:

```
IndexStatsIT > testConcurrentIndexingAndStatsRequests {p0={"cluster.indices.replication.strategy":"SEGMENT"}} FAILED
    java.lang.AssertionError:
    Expected: an empty collection
         but: <[[test][3] failed, reason [BroadcastShardOperationFailedException[operation indices:monitor/stats failed]; nested: NullPointerException[Cannot invoke "java.util.Map$Entry.getValue()" because "highestEntry" is null]; ]]>
        at __randomizedtesting.SeedInfo.seed([B56C9D0503BD16AE:2FDF6CC817B0190]:0)
        at org.hamcrest.MatcherAssert.assertThat(MatcherAssert.java:18)
        at org.junit.Assert.assertThat(Assert.java:964)
        at org.junit.Assert.assertThat(Assert.java:930)
        at org.opensearch.indices.stats.IndexStatsIT.testConcurrentIndexingAndStatsRequests(IndexStatsIT.java:1451)
```

The `isEmpty()` check on the concurrent map is not sufficient because
entries can be removed after the check but before retrieving them.

Signed-off-by: Andrew Ross <andrross@amazon.com>

Author: andrross

CHANGELOG.md

@@ -10,6 +10,7 @@ The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 
 ### Fixed
 - Fix unnecessary refreshes on update preparation failures ([#15261](https://github.com/opensearch-project/OpenSearch/issues/15261))
+- Fix possible NullPointerException in segment replicator ([]())
 
 ### Dependencies
 

server/src/main/java/org/opensearch/indices/replication/SegmentReplicator.java

@@ -160,13 +160,17 @@ void startMergedSegmentReplication(final MergedSegmentReplicationTarget target,
      */
     public ReplicationStats getSegmentReplicationStats(final ShardId shardId) {
         final ConcurrentNavigableMap<Long, ReplicationCheckpointStats> existingCheckpointStats = replicationCheckpointStats.get(shardId);
-        if (existingCheckpointStats == null || existingCheckpointStats.isEmpty()) {
+        if (existingCheckpointStats == null) {
             return ReplicationStats.empty();
         }
 
         Map.Entry<Long, ReplicationCheckpointStats> lowestEntry = existingCheckpointStats.firstEntry();
         Map.Entry<Long, ReplicationCheckpointStats> highestEntry = existingCheckpointStats.lastEntry();
 
+        if (lowestEntry == null || highestEntry == null) {
+            return ReplicationStats.empty();
+        }
+
         long bytesBehind = highestEntry.getValue().getBytesBehind();
         long replicationLag = bytesBehind > 0L
             ? Duration.ofNanos(DateUtils.toLong(Instant.now()) - lowestEntry.getValue().getTimestamp()).toMillis()