SOLUTION · MARIADB CDC

MariaDB change data capture and real-time data replication

Real-time MariaDB CDC from the binary log, with a short list of server settings and grants your DBA can review before approving anything.

Gluesync by MOLO17 provides MariaDB CDC by reading changes from the binary log (Binlog) with a dedicated source agent that connects through a built-in JDBC driver. The agent can attach to the master, a replica, or a MaxScale instance, and it delivers inserts, updates, and deletes continuously to the databases, warehouses, lakes, and event streams your teams already use. Manage every pipeline from the Core Hub web UI, the Gluesync control plane.

Start a Gluesync trial Talk to us

VENDOR COMPATIBILITY

Battle-tested on every MariaDB you run

The same Gluesync MariaDB agent is tested against each vendor offering below, self-managed or fully managed, so capture behaves the same wherever MariaDB runs.

MariaDB Source Target Tested
Amazon RDS for MariaDB Source Target Tested

WHO THIS IS FOR

MariaDB teams moving operational data into analytics and modern platforms

MariaDB DBAs who must approve the change before it reaches production: binlog_format=ROW with the binary log enabled, a dedicated user with REPLICATION SLAVE, read access to the captured tables, and a staging schema
Data platform leads feeding Snowflake, BigQuery, data lakes, Kafka, MongoDB, or Couchbase from MariaDB systems of record, who want one Core Hub for sources and targets instead of a separate Kafka Connect deployment to run and monitor
Engineers who prototyped Debezium or a DIY binlog reader and now own Kafka Connect, offsets, schema history, and on-call for a pipeline the business treats as infrastructure
Teams comparing Qlik Replicate, Fivetran HVR, or AWS DMS for MariaDB sources who want commercial binlog capture with best-in-class enterprise support, rated 4.9/5 by customers and heterogeneous targets

THE PROBLEM

MariaDB data that moves only in batches goes stale

MariaDB is often the system of record, so every nightly dump, scheduled extract, or timestamp-based poll leaves analytics and downstream systems working from a copy that is already behind. Polling queries also compete with the transactional workload for CPU and I/O on the server. Buyers searching for MariaDB CDC or MariaDB replication usually weigh Debezium with Kafka Connect, MaxScale-based pipelines, DIY binlog readers, and commercial tools such as Qlik Replicate, Fivetran HVR, or AWS DMS.

Gluesync addresses that pattern with per-agent CDC. The MariaDB source agent reads the binary log, and target agents write to the destination you choose. Each supported system has its own agent and native change stream, so one Core Hub runs the pipeline from MariaDB to every target.

HOW IT WORKS

How Gluesync does MariaDB CDC

Binary log streaming

The MariaDB agent connects through a built-in JDBC driver and reads Binlog events from a master, a replica, or a MaxScale instance over TCP sockets, depending on how you configure it. Source support starts at MariaDB 10.11. The agent also works as a target, so the same MariaDB can feed other systems and receive changes from them. MariaDB agent docs ↗

Row changes: inserts, updates, and deletes are read from the binary log continuously, for near real-time replication.
TRUNCATE: read from the binary log and forwarded to the target as a truncate, not as a stream of deletes.
Before and after images: enabled automatically in this agent, so only the changed values are applied to the target.

What your DBA will be asked to enable

The agent needs row-based binary logging and a dedicated user with the grants below. Full statements are in the setup guide.

Row-based binary logging: log_bin is enabled and binlog_format=ROW is set in the server configuration file, followed by a MariaDB restart.
Binlog retention: set expire_logs_days against the longest pause you plan for.
Least-privilege user: REPLICATION SLAVE, SELECT on mysql.innodb_table_stats, and read access to the source tables, plus create and write rights in a staging schema.
Managed services: for AWS RDS and other DBaaS, set Use SSL to true during agent setup, because those services accept only SSL connections.

Connection and topology

Where to attach: the master, a replica, or a MaxScale instance, as your configuration requires.
TLS: Enable TLS is optional and defaults to off. Provide a certificate path when your server requires one.
Setup: the Core Hub web UI, or the REST credentials endpoint, with the default port 3306.

Resilience and recovery

The source agent buffers changes in ArenaCache, the local source-side cache used by Gluesync agents, so a Core Hub restart or a short network outage does not force a resynchronization. Cache-based readers, MariaDB among them, persist their position in Core Hub rather than in a local checkpoint file, so moving or restarting an agent does not reset the reader. Prometheus series expose the binlog position and operation sequence for MariaDB readers.

Architecture around Core Hub

Lightweight agents sit close to each system; Core Hub hosts the web UI and REST APIs and routes changes between source and target agents. A pipeline groups the MariaDB source agent, its target agents, and the entities they replicate. The MariaDB agent supports snapshots for initial loads and ad-hoc refreshes, bulk load on the target side, and both the source and target roles. Core Hub and agents deploy with Docker, Docker Compose, or Kubernetes.

Explore the general CDC streaming architecture →

CAPTURE OPTIONS

MariaDB source capture: one agent, one technique

Gluesync has one MariaDB source agent. It reads the binary log, so the decisions that matter are where it connects (master, replica, or MaxScale) and how the server is configured, not which capture method to pick.

Agent	Capture technique	Versions	Best for
MariaDB agent ↗	Binary log (Binlog) streaming over TCP from a master, replica, or MaxScale instance	MariaDB 10.11 and later for the source role; any MariaDB version for the target role	Any MariaDB estate where row-based binary logging can be enabled and a dedicated user can be granted the privileges above. Includes Amazon RDS for MariaDB and other DBaaS deployments with SSL enabled.

TARGETS AND TOPOLOGIES

Keep MariaDB as the system of record, or as the destination

Use the integrations directory to pair MariaDB as source or target with relational engines, NoSQL stores, cloud warehouses, object and lake storage, or event streams, subject to each agent's documented source and target role. The MariaDB target agent writes over the built-in JDBC driver with TLS, and honors the Core Hub TRUNCATE before snapshot option when enabled.

Gluesync keeps pace with your change volume at any scale. MOLO17 Professional Services can size the deployment with your team.

Target agents write in optimized batches, never row by row, and switch to native bulk load for both snapshots and CDC on targets such as Snowflake, Google BigQuery, Amazon Redshift, Microsoft SQL Server, and PostgreSQL.

Offload reporting and API reads from MariaDB to another database or cache: see database offload
Feed Snowflake or BigQuery continuously from MariaDB: see warehouse sync
Publish MariaDB changes to Apache Kafka or another event-streaming platform for microservices and integration
Run a snapshot plus CDC to keep MariaDB and a cloud target aligned until cutover: see cloud migration

FAIR, HIGH-LEVEL COMPARISON

Where Gluesync fits among MariaDB CDC approaches

Approach	What buyers usually get	Where Gluesync fits
MariaDB native replication	Built into the server and suited to MariaDB-to-MariaDB copies; delivering to warehouses, NoSQL stores, or streams needs other tooling	Reads the same binary log but delivers to heterogeneous targets under one Core Hub, and can write back into MariaDB as a target
Debezium with Kafka Connect	Open-source binlog reader; you run Kafka, Kafka Connect, offsets, schema history, sinks, and upgrades	Productized binlog agent with Core Hub operations and MOLO17 enterprise support; read the Debezium alternative comparison
MaxScale-based change streaming	Keeps the stream inside MariaDB's proxy layer; routing, consumers, and downstream delivery are yours to build and run	Gluesync can attach through a MaxScale node as a documented connection point, then deliver to the targets you choose
Qlik Replicate, Fivetran HVR, or AWS DMS class CDC	Mature replication portfolios with broad source coverage; packaging and support terms vary by product and cloud	Agent-based binlog capture with heterogeneous targets under one Core Hub; see migrating to Gluesync for supported paths
DIY binlog readers and timestamp polling	Full control with high build and operations cost; polling queries compete with production load	Productized agent with snapshots, positions persisted in Core Hub, recovery, and monitoring; read the MariaDB agent launch post

MariaDB CDC research and implementation detail

What is MariaDB CDC with Gluesync?

Change data capture from MariaDB by reading row changes from the binary log (Binlog) through the MariaDB agent, then delivering inserts, updates, and deletes to configured targets via Gluesync agents and Core Hub. TRUNCATE operations are captured as well.

Which MariaDB versions does the agent support?

As a source, the MariaDB agent captures from MariaDB 10.11 and later. As a target, it writes to any MariaDB version over the built-in JDBC driver.

What must be enabled on the MariaDB server?

Binary logging with binlog_format=ROW, set in the server configuration file followed by a restart. A dedicated user needs REPLICATION SLAVE, SELECT on mysql.innodb_table_stats, read access to the captured tables, and create and write rights in a staging schema.

Can the agent read from a replica or a MaxScale instance?

Yes. The agent can connect to the master, a replica, or a MaxScale node, depending on your configuration.

Does Gluesync support Amazon RDS for MariaDB?

Yes. Amazon RDS for MariaDB is supported. Set Use SSL to true during agent setup, because RDS accepts only SSL connections.

Can Gluesync write to MariaDB, not only read from it?

Yes. The MariaDB target agent supports bulk load through staging tables, which needs local_infile=1 on the server, and the Core Hub TRUNCATE before snapshot option.

Does the agent support initial snapshots?

Yes. The agent works with Core Hub snapshot tasks for initial loads and ad-hoc refreshes, so you can seed a target and then keep it current from the binary log.

CDC BY SOURCE DATABASE

Other sources Gluesync captures from

Evaluate Gluesync with your MariaDB binary logs

Start a trial on your infrastructure, or talk to MOLO17 about the right MariaDB setup for your binary log configuration, prerequisites, and targets.

Start a Gluesync trial Talk to MOLO17