[libbeat][parquet reader] - Added debug logs & improved batch_size tracking

CHANGELOG-developer.next.asciidoc

@@ -202,6 +202,7 @@ The list below covers the major changes between 7.0.0-rc2 and main only.
 - Bump version of elastic/toutoumomoma to remove internal forks of stdlib debug packages. {pull}40325[40325]
 - Refactor x-pack/filebeat/input/websocket for generalisation. {pull}40308[40308]
 - Add a configuration option for TCP/UDP network type. {issue}40407[40407] {pull}40623[40623]
+- Added debug logging to parquet reader in x-pack/libbeat/reader. {pull}40651[40651]
 
 ==== Deprecated
 

x-pack/libbeat/reader/parquet/parquet.go

@@ -14,65 +14,79 @@ import (
 	"github.com/apache/arrow/go/v14/parquet"
 	"github.com/apache/arrow/go/v14/parquet/file"
 	"github.com/apache/arrow/go/v14/parquet/pqarrow"
+
+	"github.com/elastic/elastic-agent-libs/logp"
 )
 
 // BufferedReader parses parquet inputs from io streams.
 type BufferedReader struct {
 	cfg          *Config
 	fileReader   *file.Reader
 	recordReader pqarrow.RecordReader
+	log          *logp.Logger
 }
 
 // NewBufferedReader creates a new reader that can decode parquet data from an io.Reader.
 // It will return an error if the parquet data stream cannot be read.
 // Note: As io.ReadAll is used, the entire data stream would be read into memory, so very large data streams
 // may cause memory bottleneck issues.
 func NewBufferedReader(r io.Reader, cfg *Config) (*BufferedReader, error) {
-	batchSize := 1
-	if cfg.BatchSize > 1 {
-		batchSize = cfg.BatchSize
+	log := logp.L().Named("reader.parquet")
+
+	if cfg.BatchSize == 0 {
+		cfg.BatchSize = 1
 	}
+	log.Debugw("creating parquet reader", "batch_size", cfg.BatchSize)
 
 	// reads the contents of the reader object into a byte slice
 	data, err := io.ReadAll(r)
 	if err != nil {
 		return nil, fmt.Errorf("failed to read data from stream reader: %w", err)
 	}
+	log.Debugw("read data from stream reader", "size", len(data))
 
 	// defines a memory allocator for allocating memory for Arrow objects
 	pool := memory.NewCheckedAllocator(&memory.GoAllocator{})
-
+	// constructs a parquet file reader object from the byte slice data
 	pf, err := file.NewParquetReader(bytes.NewReader(data), file.WithReadProps(parquet.NewReaderProperties(pool)))
 	if err != nil {
 		return nil, fmt.Errorf("failed to create parquet reader: %w", err)
 	}
+	log.Debugw("created parquet reader")
 
 	// constructs a reader for converting to Arrow objects from an existing parquet file reader object
 	reader, err := pqarrow.NewFileReader(pf, pqarrow.ArrowReadProperties{
 		Parallel:  cfg.ProcessParallel,
-		BatchSize: int64(batchSize),
+		BatchSize: int64(cfg.BatchSize),
 	}, pool)
 	if err != nil {
 		return nil, fmt.Errorf("failed to create pqarrow parquet reader: %w", err)
 	}
+	log.Debugw("created pqarrow parquet reader")
 
 	// constructs a record reader that is capable of reding entire sets of arrow records
 	rr, err := reader.GetRecordReader(context.Background(), nil, nil)
 	if err != nil {
 		return nil, fmt.Errorf("failed to create parquet record reader: %w", err)
 	}
+	log.Debugw("initialization process completed")
 
 	return &BufferedReader{
 		cfg:          cfg,
 		recordReader: rr,
 		fileReader:   pf,
+		log:          log,
 	}, nil
 }
 
 // Next advances the pointer to point to the next record and returns true if the next record exists.
 // It will return false if there are no more records to read.
 func (sr *BufferedReader) Next() bool {
-	return sr.recordReader.Next()
+	next := sr.recordReader.Next()
+	if !next {
+		sr.log.Debugw("no more records to read", "next", next)
+	}
+	return next
 }
 
 // Record reads the current record from the parquet file and returns it as a JSON marshaled byte slice.
@@ -81,13 +95,16 @@ func (sr *BufferedReader) Next() bool {
 func (sr *BufferedReader) Record() ([]byte, error) {
 	rec := sr.recordReader.Record()
 	if rec == nil {
+		sr.log.Debugw("reached the end of the record reader", "record_reader", rec)
 		return nil, io.EOF
 	}
 	defer rec.Release()
 	val, err := rec.MarshalJSON()
 	if err != nil {
 		return nil, fmt.Errorf("failed to marshal JSON for parquet value: %w", err)
 	}
+	sr.log.Debugw("records successfully read", "batch_size", sr.cfg.BatchSize)
+
 	return val, nil
 }
 

x-pack/libbeat/reader/parquet/parquet_test.go

@@ -19,6 +19,8 @@ import (
 	"github.com/apache/arrow/go/v14/arrow/memory"
 	"github.com/apache/arrow/go/v14/parquet/pqarrow"
 	"github.com/stretchr/testify/assert"
+
+	"github.com/elastic/elastic-agent-libs/logp"
 )
 
 // all test files are read from/stored within the "testdata" directory
@@ -55,6 +57,7 @@ func TestParquetWithRandomData(t *testing.T) {
 		},
 	}
 
+	logp.TestingSetup()
 	for i, tc := range testCases {
 		name := fmt.Sprintf("Test parquet files with rows=%d, and columns=%d", tc.rows, tc.columns)
 		t.Run(name, func(t *testing.T) {
@@ -189,6 +192,7 @@ func TestParquetWithFiles(t *testing.T) {
 		},
 	}
 
+	logp.TestingSetup()
 	for _, tc := range testCases {
 		name := fmt.Sprintf("Test parquet files with source file=%s, and target comparison file=%s", tc.parquetFile, tc.jsonFile)
 		t.Run(name, func(t *testing.T) {