add a ms instead of ns to end time of the rebound chunk interval (#5351)

* add a ms instead of ns to end time of the rebound chunk interval

conversion between ms and ns loses ns precision which causes us to not include all the logs

* improve error when failing to rebound chunk

pkg/chunkenc/memchunk.go

@@ -911,8 +911,8 @@ func (c *MemChunk) Blocks(mintT, maxtT time.Time) []Block {
 
 // Rebound builds a smaller chunk with logs having timestamp from start and end(both inclusive)
 func (c *MemChunk) Rebound(start, end time.Time) (Chunk, error) {
-	// add a nanosecond to end time because the Chunk.Iterator considers end time to be non-inclusive.
-	itr, err := c.Iterator(context.Background(), start, end.Add(time.Nanosecond), logproto.FORWARD, log.NewNoopPipeline().ForStream(labels.Labels{}))
+	// add a millisecond to end time because the Chunk.Iterator considers end time to be non-inclusive.
+	itr, err := c.Iterator(context.Background(), start, end.Add(time.Millisecond), logproto.FORWARD, log.NewNoopPipeline().ForStream(labels.Labels{}))
 	if err != nil {
 		return nil, err
 	}

pkg/storage/stores/shipper/compactor/retention/retention.go

@@ -152,7 +152,7 @@ func markforDelete(ctx context.Context, tableName string, marker MarkerStorageWr
 			if len(nonDeletedIntervals) > 0 {
 				wroteChunks, err := chunkRewriter.rewriteChunk(ctx, c, nonDeletedIntervals)
 				if err != nil {
-					return false, false, err
+					return false, false, fmt.Errorf("failed to rewrite chunk %s for interval %s with error %s", c.ChunkID, nonDeletedIntervals, err)
 				}
 
 				if wroteChunks {