go-pq-cdc-kafka

Kafka connector for go-pq-cdc.

go-pq-cdc-kafka is a high-performance tool designed to replicate PostgreSQL data to Kafka using PostgreSQL's replication protocol. It provides an efficient alternative to Debezium, particularly in resource-constrained environments.

Key Features

• Optimized for Speed and Efficiency: Minimal resource consumption and faster processing, designed to handle high-throughput data replication.
• Real-Time Data Streaming: Streams data directly from PostgreSQL to Kafka, ensuring up-to-date synchronization across systems.
• Initial Snapshot Support: Capture existing data before starting CDC, ensuring downstream systems receive both historical and real-time data.
• Automatic Failover: In the event of a failure, go-pq-cdc-kafka can quickly recover and resume data replication.
• Concurrency: Built with Go's concurrency model (goroutines and channels), ensuring lightweight and highly performant parallel operations.

Why Choose go-pq-cdc-kafka Over Debezium?

Debezium vs go-pq-cdc-kafka benchmark

go-pq-cdc-kafka significantly outperforms Debezium in terms of speed and resource efficiency, particularly for PostgreSQL databases. Here’s a comparison:

Benchmark Results

	go-pq-cdc-kafka	Debezium
Row Count	10 million	10 million
Elapsed Time	2.5 minutes	21 minutes
CPU Usage (Max)	44%	181%
Memory Usage (Max)	130 MB	1.07 GB
Received Traffic	4.36 MiB/s	7.97 MiB/s
Sent Traffic	5.96 MiB/s	6.27 MiB/s

Availability and Failover

The go-pq-cdc-kafka ensures high availability with passive/active modes for PostgreSQL Change Data Capture (CDC).

• Active Mode: When the PostgreSQL replication slot is active, go-pq-cdc-kafka continuously monitors changes and streams them to Kafka.

• Passive Mode: If the replication slot becomes inactive, it automatically captures the slot and resumes data streaming. Additionally, other deployments monitor the slot’s status, ensuring redundancy and failover capabilities.

This architecture guarantees minimal downtime and continuous data synchronization, even in the event of failure. Additionally, Go's faster cold starts provide quicker recovery times compared to Debezium, further minimizing potential downtime.

📸 NEW: Snapshot Feature

Capture existing data before starting CDC! The snapshot feature enables initial data synchronization, ensuring downstream systems (Kafka) receive both historical and real-time data.

✨ Key Highlights:

• Zero Data Loss: Consistent point-in-time snapshot using PostgreSQL's pg_export_snapshot()
• Chunk-Based Processing: Memory-efficient processing of large tables
• Multi-Instance Support: Parallel processing across multiple instances for faster snapshots
• Crash Recovery: Automatic resume from failures with chunk-level tracking
• No Duplicates: Seamless transition from snapshot to CDC mode
• Flexible Modes: Choose between initial, never, or snapshot_only based on your needs

Snapshot Modes

Mode	Description	Use Case
initial	Takes snapshot only if no previous snapshot exists, then starts CDC	First-time setup with existing data
never	Skips snapshot entirely, starts CDC immediately	New tables or when historical data not needed
snapshot_only	Takes snapshot and exits (no CDC, no replication slot required)	One-time data migration or backfill

How It Works

1. Snapshot Phase: Captures existing data in chunks for memory efficiency
2. Consistent Point: Uses PostgreSQL's pg_export_snapshot() to ensure data consistency
3. CDC Phase: Seamlessly transitions to real-time change data capture
4. No Gaps: Ensures all changes during snapshot are captured via CDC

Identifying Snapshot vs CDC Messages

Your handler function can distinguish between snapshot and CDC messages:

func Handler(msg *cdc.Message) []gokafka.Message {
    // Check if this is a snapshot message (historical data)
    if msg.Type.IsSnapshot() {
        // Handle snapshot data
        return []gokafka.Message{{
            Headers: []gokafka.Header{
                {Key: "operation", Value: []byte("SNAPSHOT")},
                {Key: "source", Value: []byte("initial-snapshot")},
            },
            Key:   key,
            Value: data,
        }}
    }
    
    // Handle real-time CDC operations
    if msg.Type.IsInsert() { /* ... */ }
    if msg.Type.IsUpdate() { /* ... */ }
    if msg.Type.IsDelete() { /* ... */ }
}

For detailed configuration and usage, see the snapshot example.

Usage

⚠️ For production usage check the production tutorial doc

⚠️ For other usages check the dockerfile and code at examples.

go get github.com/Trendyol/go-pq-cdc-kafka

package main

import (
	"context"
	"encoding/json"
	"log/slog"
	"os"
	"strconv"
	"time"

	cdc "github.com/Trendyol/go-pq-cdc-kafka"
	"github.com/Trendyol/go-pq-cdc-kafka/config"
	cdcconfig "github.com/Trendyol/go-pq-cdc/config"
	"github.com/Trendyol/go-pq-cdc/pq/publication"
	"github.com/Trendyol/go-pq-cdc/pq/slot"
	gokafka "github.com/segmentio/kafka-go"
)

func main() {
	slog.SetDefault(slog.New(slog.NewJSONHandler(os.Stdout, nil)))
	ctx := context.TODO()
	cfg := config.Connector{
		CDC: cdcconfig.Config{
			Host:      "127.0.0.1",
			Username:  "cdc_user",
			Password:  "cdc_pass",
			Database:  "cdc_db",
			DebugMode: false,
			Publication: publication.Config{
				CreateIfNotExists: true,
				Name:              "cdc_publication",
				Operations: publication.Operations{
					publication.OperationInsert,
					publication.OperationDelete,
					publication.OperationTruncate,
					publication.OperationUpdate,
				},
				Tables: publication.Tables{publication.Table{
					Name:            "users",
					ReplicaIdentity: publication.ReplicaIdentityFull,
				}},
			},
			Slot: slot.Config{
				CreateIfNotExists:           true,
				Name:                        "cdc_slot",
				SlotActivityCheckerInterval: 3000,
			},
			Metric: cdcconfig.MetricConfig{
				Port: 8081,
			},
			Logger: cdcconfig.LoggerConfig{
				LogLevel: slog.LevelInfo,
			},
		},
		Kafka: config.Kafka{
			TableTopicMapping:           map[string]string{"public.users": "users.0"},
			Brokers:                     []string{"localhost:19092"},
			AllowAutoTopicCreation:      true,
			ProducerBatchTickerDuration: time.Millisecond * 200,
		},
	}

	connector, err := cdc.NewConnector(ctx, cfg, Handler)
	if err != nil {
		slog.Error("new connector", "error", err)
		os.Exit(1)
	}

	defer connector.Close()
	connector.Start(ctx)
}

func Handler(msg *cdc.Message) []gokafka.Message {
	slog.Info("change captured", "message", msg)
	if msg.Type.IsUpdate() || msg.Type.IsInsert() {
		msg.NewData["operation"] = msg.Type
		newData, _ := json.Marshal(msg.NewData)

		return []gokafka.Message{
			{
				Headers: nil,
				Key:     []byte(strconv.Itoa(int(msg.NewData["id"].(int32)))),
				Value:   newData,
			},
		}
	}

	if msg.Type.IsDelete() {
		msg.OldData["operation"] = msg.Type
		oldData, _ := json.Marshal(msg.OldData)

		return []gokafka.Message{
			{
				Headers: nil,
				Key:     []byte(strconv.Itoa(int(msg.OldData["id"].(int32)))),
				Value:   oldData,
			},
		}
	}

	return []gokafka.Message{}
}

Examples

• Simple
• Snapshot

Configuration

Variable	Type	Required	Default	Description	Options
cdc.host	string	yes	-	PostgreSQL host	Should be a valid hostname or IP address. Example: localhost.
cdc.username	string	yes	-	PostgreSQL username	Should have sufficient privileges to perform required database operations.
cdc.password	string	yes	-	PostgreSQL password	Keep secure and avoid hardcoding in the source code.
cdc.database	string	yes	-	PostgreSQL database	The database must exist and be accessible by the specified user.
cdc.debugMode	bool	no	false	For debugging purposes	Enables pprof for trace.
cdc.metric.port	int	no	8080	Set API port	Choose a port that is not in use by other applications.
cdc.logger.logLevel	string	no	info	Set logging level	[DEBUG, WARN, INFO, ERROR]
cdc.logger.logger	Logger	no	slog	Set logger	Can be customized with other logging frameworks if slog is not used.
cdc.publication.createIfNotExists	bool	no	-	Create publication if not exists. Otherwise, return publication is not exists error.
cdc.publication.name	string	yes	-	Set PostgreSQL publication name	Should be unique within the database.
cdc.publication.operations	[]string	yes	-	Set PostgreSQL publication operations. List of operations to track; all or a subset can be specified.	INSERT: Track insert operations. UPDATE: Track update operations. DELETE: Track delete operations.
cdc.publication.tables	[]Table	yes	-	Set tables which are tracked by data change capture	Define multiple tables as needed.
cdc.publication.tables[i].name	string	yes	-	Set the data change captured table name	Must be a valid table name in the specified database.
cdc.publication.tables[i].replicaIdentity	string	yes	-	Set the data change captured table replica identity [FULL, DEFAULT]	FULL: Captures all columns of old row when a row is updated or deleted. DEFAULT: Captures only the primary key of old row when a row is updated or deleted.
publication.tables[i].schema	string	no	public	Set the data change captured table schema name	Must be a valid table name in the specified database.
cdc.slot.createIfNotExists	bool	no	-	Create replication slot if not exists. Otherwise, return replication slot is not exists error.
cdc.slot.name	string	yes	-	Set the logical replication slot name	Should be unique and descriptive.
cdc.slot.slotActivityCheckerInterval	int	yes	1000	Set the slot activity check interval time in milliseconds	Specify as an integer value in milliseconds (e.g., 1000 for 1 second).
cdc.snapshot.enabled	bool	no	false	Enable initial snapshot feature	When enabled, captures existing data before starting CDC.
cdc.snapshot.mode	string	no	never	Snapshot mode: initial, never, or snapshot_only	initial: Take snapshot only if no previous snapshot exists, then start CDC. never: Skip snapshot, start CDC immediately. snapshot_only: Take snapshot and exit (no CDC).
cdc.snapshot.chunkSize	int64	no	8000	Number of rows per chunk during snapshot	Adjust based on table size. Larger chunks = fewer chunks but more memory per chunk.
cdc.snapshot.claimTimeout	time.Duration	no	30s	Timeout to reclaim stale chunks	If a worker doesn't send heartbeat for this duration, chunk is reclaimed by another worker.
cdc.snapshot.heartbeatInterval	time.Duration	no	5s	Interval for worker heartbeat updates	Workers send heartbeat every N seconds to indicate they're processing a chunk.
cdc.snapshot.instanceId	string	no	auto	Custom instance identifier (optional)	Auto-generated as hostname-pid if not specified. Useful for tracking workers in multi-instance scenarios.
cdc.snapshot.tables	[]Table	no*	-	Tables to snapshot (required for snapshot_only mode, optional for initial mode)	snapshot_only: Must be specified here (independent from publication). initial: If specified, must be a subset of publication tables. If not specified, all publication tables are snapshotted.
kafka.tableTopicMapping	map[string]string	yes	-	Mapping of PostgreSQL table events to Kafka topics	Maps table names to Kafka topics.
kafka.brokers	[]string	yes	-	Broker IP and port information
kafka.producerBatchSize	integer	no	2000	Maximum message count for batch, if exceeded, flush will be triggered.
kafka.producerBatchBytes	string	no	10mb	Maximum size (bytes) for batch. If exceeded, flush will be triggered.
kafka.producerMaxAttempts	int	no	maxInt	Limit on how many attempts will be made to deliver a message.
kafka.producerBatchTickerDuration	time.Duration	no	10s	Batch is flushed automatically at specific time intervals for long-waiting messages in the batch.
kafka.readTimeout	time.Duration	no	30s	Timeout for read operations in segmentio/kafka-go.
kafka.writeTimeout	time.Duration	no	30s	Timeout for write operations in segmentio/kafka-go.
kafka.compression	integer	no	0	Compression can be used if message size is large, CPU usage may be affected.	0: None 1: Gzip 2: Snappy 3: Lz4 4: Zstd
kafka.balancer	string	no	Hash	Define balancer strategy.	[Hash, LeastBytes, RoundRobin, ReferenceHash, CRC32Balancer, Murmur2Balancer]
kafka.requiredAcks	integer	no	1	Number of acknowledgments from partition replicas required before receiving a response.	0: fire-and-forget 1: wait for leader to acknowledge writes -1: wait for full ISR
kafka.secureConnection	bool	no	false	Enable secure Kafka connection.
kafka.rootCA	[]byte	no	-	Define root CA certificate.
kafka.interCA	[]byte	no	-	Define intermediate CA certificate.
kafka.scramUsername	string	no	-	Define SCRAM username.
kafka.scramPassword	string	no	-	Define SCRAM password.
kafka.metadataTTL	time.Duration	no	60s	TTL for the metadata cached by segmentio/kafka-go. Increase it to reduce network requests.	For more detail, check docs.
kafka.metadataTopics	[]string	no	-	Topic names for the metadata cached by segmentio/kafka-go. Define topics here that the connector may produce.	For more detail, check docs.
kafka.clientID	string	no	-	Unique identifier that the transport communicates to the brokers.	For more detail, check docs.
kafka.allowAutoTopicCreation	bool	no	false	Create topic if missing.	For more detail, check docs.

API

Endpoint	Description
GET /status	Returns a 200 OK status if the client is able to ping the PostgreSQL server successfully.
GET /metrics	Prometheus metric endpoint.
GET /debug/pprof/*	(Only for debugMode=true) pprof

Exposed Metrics

The client collects relevant metrics related to PostgreSQL change data capture (CDC) and makes them available at the /metrics endpoint.

Metric Name	Description	Labels	Value Type
go_pq_cdc_kafka_process_latency_current	The latest kafka connector process latency in nanoseconds.	slot_name, host	Gauge
go_pq_cdc_kafka_bulk_request_process_latency_current	The latest kafka connector bulk request process latency in nanoseconds.	slot_name, host	Gauge
go_pq_cdc_kafka_write_total	The total number of successful in write operation to kafka.	slot_name, host, topic_name	Counter
go_pq_cdc_kafka_err_total	The total number of unsuccessful in write operation to kafka.	slot_name, host, topic_name	Counter

Snapshot Metrics

Metric Name	Description	Labels	Value Type
go_pq_cdc_snapshot_in_progress	Indicates whether snapshot is currently in progress (1 for active, 0 for inactive).	slot_name, host	Gauge
go_pq_cdc_snapshot_total_tables	Total number of tables to snapshot.	slot_name, host	Gauge
go_pq_cdc_snapshot_total_chunks	Total number of chunks to process across all tables.	slot_name, host	Gauge
go_pq_cdc_snapshot_completed_chunks	Number of chunks completed in snapshot.	slot_name, host	Gauge
go_pq_cdc_snapshot_total_rows	Total number of rows read during snapshot.	slot_name, host	Counter
go_pq_cdc_snapshot_duration_seconds	Duration of the last snapshot operation in seconds.	slot_name, host	Gauge

You can also use all cdc related metrics explained here. All cdc related metrics are automatically injected. It means you don't need to do anything.

Compatibility

go-pq-cdc-kafka Version	Minimum PostgreSQL Server Version
0.0.0 or higher	14

Breaking Changes

Date taking effect	Version	Change	How to check
-	-	-	-