SOLUTION · AZURE COSMOS DB CDC

Azure Cosmos DB change data capture and real-time data replication

Change feed capture from Azure Cosmos DB, without writing and running your own change feed consumer.

Gluesync by MOLO17 provides Azure Cosmos DB CDC through a dedicated source agent that reads the change feed of your monitored container and delivers changes 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.

WHO THIS IS FOR

Azure teams moving application data from Cosmos DB into analytics and operations

  • Azure platform admins and DBAs who must approve a change feed consumer, with the hostname, Azure Key, monitored container, and lease container written down before anything runs, and TLS on port 443 confirmed for the network path
  • Data platform leads feeding Snowflake, SQL Server, PostgreSQL, data lakes, or Kafka from Cosmos DB, who want one Core Hub for every source and target instead of a separate pipeline per engine
  • Application engineers who wrote change feed consumers or Azure Functions triggers and now own leases, retries, and monitoring for a pipeline the business treats as infrastructure
  • Teams replacing Kafka Connect pipelines that read Cosmos DB, or weighing a commercial path with best-in-class enterprise support, rated 4.9/5 by customers and heterogeneous targets against building and operating their own consumer

THE PROBLEM

Cosmos DB data that only moves in batches goes stale, and the DIY paths are expensive to run

Cosmos DB often holds the live data of an application, so scheduled exports leave analytics and downstream services working from copies that are already behind. Each full scan also adds load to the same container the application uses. Buyers searching for Cosmos DB CDC or Cosmos DB replication usually compare three options: custom change feed code, Azure Functions triggers, and Kafka Connect pipelines. Each one leaves the team owning leases, retries, and the sinks.

Gluesync addresses that pattern with per-agent CDC. The Azure Cosmos DB agent reads the change feed of the monitored container and delivers changes to the targets you choose. Each supported system has its own agent and native change stream, and all of them run under the same Core Hub.

HOW IT WORKS

How Gluesync does Azure Cosmos DB CDC

Native change feed capture

The agent implements the Azure Cosmos DB change feed processor, which is part of the Azure V4 SDK, and it embeds the official azure-cosmos driver. Changes are read from the monitored container.

  • Inserts and updates to the monitored container are reflected in the change feed and captured.
  • Leases track progress. The change feed processor stores its state in a lease container, so processing resumes where it left off and can be coordinated across workers.
  • Snapshots for the first load. Full snapshot tasks seed the target before CDC takes over. Change data capture docs ↗

What to prepare before the first run

Two things must be in place before the agent starts.

  1. Lease container, prepared up front. The lease container stores the state of the change feed processor and coordinates processing across workers. Create it in the same database with the partition key path /id, which is the default, using the Azure CLI command in the setup guide. Use the Advanced settings tab if you need a different partition key.
  2. Service key and TLS. Gluesync connects with the Azure Cosmos DB service key, entered as the Azure Key, over HTTPS to port 443 with TLS. The network path needs outbound access on that port.

Regions and connectivity

  • Regions. The agent works with Azure Cosmos DB DBaaS in any region.
  • Connections. You enter the hostname, the port (default 443), the database name, the Azure Key, the monitored container, and the lease container name. Connections use TLS.

Architecture around Core Hub

Agents run close to the systems they read and write, and Core Hub coordinates them: it hosts the web UI and REST APIs and routes every change from the Cosmos DB source agent to its targets. A pipeline groups the source agent, its target agents, and the entities (containers) they replicate. The agent also supports full snapshot tasks that seed a Cosmos DB container before CDC starts, and it supports the target role, so Cosmos DB can be the destination as well as the origin. Core Hub and agents deploy with Docker, Docker Compose, or Kubernetes.

Explore the general CDC streaming architecture →

TARGETS AND TOPOLOGIES

Keep Cosmos DB as the system of record while modernizing destinations

Use the integrations directory to pair Azure Cosmos DB 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 Azure Cosmos DB agent supports the target role through the Azure Cosmos Java SDK over TLS, so Cosmos DB can also receive data from other sources.

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 analytics and reporting reads from Cosmos DB to a relational or NoSQL target: see database offload
  • Feed Snowflake, a data lake, or a cloud warehouse continuously: see warehouse sync and data lake
  • Publish Cosmos DB changes to Apache Kafka or Google Pub/Sub for microservices and integration
  • Seed a Cosmos DB container from another source with a snapshot, then keep it aligned until cutover: see cloud migration

FAIR, HIGH-LEVEL COMPARISON

Where Gluesync fits among Cosmos DB change feed approaches

ApproachWhat buyers usually getWhere Gluesync fits
Custom change feed code or an Azure Functions trigger Direct access to the change feed in your own code; you build and run the consumer, lease handling, retries, and each sink A packaged source agent with documented lease container setup, snapshots, and one Core Hub for every target; see CDC streaming
Azure Synapse Link or Microsoft Fabric mirroring Analytics copies inside the Microsoft analytics stack; the destination is tied to that stack Delivers to targets beyond one analytics stack, such as Snowflake, Kafka, or relational databases, under one Core Hub
Kafka Connect source connector for Cosmos DB Change feed events in Kafka topics; you run Kafka Connect, its offsets, and the sinks that read the topics Direct delivery to Snowflake, SQL databases, and other targets without a Kafka cluster in that path; Kafka stays a supported target, see integrations
Azure Data Factory or scheduled copy jobs Scheduled batch copies that suit periodic loads; each run moves what changed since the last window Continuous capture of inserts and updates, with snapshots for the first load; see warehouse sync
DIY consumers and sink jobs Full control; you own the leases, retries, monitoring, and every sink A productized agent with documented setup, Core Hub monitoring, and MOLO17 enterprise support, so the team runs one pipeline model instead of a custom service

FAQ

Azure Cosmos DB CDC questions

What is Azure Cosmos DB CDC with Gluesync?

Change data capture from Azure Cosmos DB through the change feed processor in the Azure V4 SDK. The source agent captures inserts and updates from the monitored container and delivers the changes to configured targets through Core Hub.

Which Azure Cosmos DB API does the agent support?

The agent connects through the Azure Cosmos Java SDK to the API for NoSQL, formerly called the SQL API.

What has to exist in Cosmos DB before the first run?

A lease container, prepared up front in the same database, with the /id partition key path by default. You also need the Azure Cosmos DB service key, the hostname, and outbound HTTPS access on port 443 with TLS.

How does the first load work?

A full snapshot task seeds the target from the monitored container, then the agent continues from the change feed so the target stays current. Core Hub shows both phases in the same pipeline.

Which regions and authentication does the agent support?

The agent works with Azure Cosmos DB DBaaS in any region and authenticates with the service key over TLS on port 443.

Can Gluesync write to Azure Cosmos DB, not only read from it?

Yes. The Azure Cosmos DB agent supports the target role, and snapshot tasks can seed a container before CDC starts. For duplicate keys on a Cosmos DB target, Core Hub can upsert, skip, or fail per entity.

Where should we start?

Start a Gluesync trial against a test account and container, then set up the lease container as the setup guide describes. MOLO17 can review your container layout and targets before anything goes to production.

Evaluate Gluesync with your Cosmos DB change feed

Start a trial on your infrastructure, or talk to MOLO17 about the lease container, snapshots, and the targets for your Azure Cosmos DB accounts.