dbmazz
What it is
dbmazz is a Rust daemon for moving PostgreSQL change events to a sink through logical replication. It streams inserts, updates, and deletes, and the project reports sub-second replication lag and about 11 MB of resident memory per daemon. PostgreSQL and MySQL are listed as sources; stable sinks include StarRocks, PostgreSQL, and Snowflake, while S3/Iceberg remains on the roadmap. A quickstart demo is available through the ez-cdc CLI, and an official Docker image can run the daemon. The CLI supports Linux and macOS on amd64 and arm64. The tool includes a terminal dashboard, Prometheus metrics, gRPC reflection, and an HTTP control plane. It can automatically create publications, replication slots, sink tables, and audit columns. Each instance handles one replication job. Self-hosting is free; the free BYOC option deploys in a customer's AWS or GCP account, with support and optional PrivateLink charged separately. The project uses Elastic License 2.0, which limits offering the software to third parties as a managed service.
For and against
For
- Free self-hosted option and free BYOC plan
- Stable sinks include StarRocks, PostgreSQL, and Snowflake
- Automatic creation of replication and sink objects
- Observability includes Prometheus metrics and a terminal dashboard
Against
- Each instance handles one replication job
- BYOC support and optional PrivateLink cost extra
- Elastic License restricts third-party managed-service offerings
- S3/Iceberg is listed only on the roadmap
The verdict
dbmazz combines PostgreSQL logical replication with several stable sink choices and automatic setup. Its per-instance job limit and licensing terms are important considerations for deployment planning.
Who it suits
It suits teams seeking PostgreSQL change-data capture without operating a streaming platform. Teams should account for the one-job-per-instance limit and the license restriction on managed-service offerings.
Plans and prices
BYOC
Deploy into your AWS or GCP account · support and optional PrivateLink connectivity are charged separately
Read 29 Sep 2026 · ez-cdc.com
Self-host dbmazz
Open-source engine · CLI and terminal dashboard · local web UI · community support
Read 29 Sep 2026 · ez-cdc.com
On the record
- Automatic setup
- The maker says ez-cdc can create publications, replication slots, sink tables, and audit columns automatically.ez-cdc.com · read 29 Sep 2026
- BYOC and data location
- EZ-CDC says BYOC runs in the customer's AWS or GCP account and that its data stays within that account.ez-cdc.com · read 29 Sep 2026
- BYOC security features
- The maker lists multi-AZ high availability, autoscaling, load balancing, optional PrivateLink, and data sovereignty for BYOC.ez-cdc.com · read 29 Sep 2026
- CLI platforms
- The project README says its CLI runs on Linux and macOS for amd64 and arm64.github.com · read 29 Sep 2026
- Intended users
- The README describes dbmazz as an option for teams seeking streaming PostgreSQL CDC without operating a streaming platform.github.com · read 29 Sep 2026
- License
- dbmazz uses Elastic License 2.0, which the README describes as allowing commercial and non-commercial use while restricting offering it to third parties as a managed service.github.com · read 29 Sep 2026
- Notable limitation
- The README says each dbmazz instance handles one replication job.github.com · read 29 Sep 2026
- Observability
- The maker lists Prometheus metrics, gRPC reflection, an HTTP control plane, and a live terminal dashboard.ez-cdc.com · read 29 Sep 2026
- Performance
- The project reports sub-second replication lag and about 11 MB of resident memory per daemon in its README.github.com · read 29 Sep 2026
- Setup
- The project provides an ez-cdc CLI quickstart demo and an official Docker image for running the daemon.github.com · read 29 Sep 2026
- Sources and sinks
- The README lists PostgreSQL and MySQL sources, with StarRocks, PostgreSQL, and Snowflake as stable sinks; S3/Iceberg is listed as roadmap.github.com · read 29 Sep 2026
- Support
- The project points users to GitHub Discussions and Issues, and its documentation lists support by email or GitHub issue.docs.ez-cdc.com · read 29 Sep 2026
- What it does
- dbmazz is a Rust daemon that reads PostgreSQL changes through logical replication and streams inserts, updates, and deletes to a sink.github.com · read 29 Sep 2026
The review
dbmazz is a self-hosted Rust daemon for capturing database changes and delivering them to supported sinks. It is best suited to teams that want PostgreSQL change-data capture without operating a separate streaming platform. Its free deployment paths and automatic setup are compelling, while the one-job-per-instance limit and Elastic License terms deserve attention before adoption.
Overview
dbmazz uses PostgreSQL logical replication to stream inserts, updates, and deletes to a sink. Initial snapshots help bootstrap a destination, and schema evolution support addresses changes to the source structure over time. The project reports sub-second replication lag and about 11 MB of resident memory per daemon. Those figures suggest a modest footprint, though they are project-reported performance claims rather than a promise for every workload.
The README names PostgreSQL and MySQL as sources, and StarRocks, PostgreSQL, and Snowflake as stable sinks. S3/Iceberg is a roadmap item, not a current stable destination. A dbmazz instance handles one replication job, so teams with multiple independent pipelines should plan for separate instances rather than assume one daemon will cover them all.
Key features
An ez-cdc CLI quickstart and an official Docker image provide direct paths to running the daemon. The maker says ez-cdc can create publications, replication slots, sink tables, and audit columns automatically. That can reduce repetitive configuration for supported setups, but does not remove the need to plan around the single-job instance limit.
Operators get Prometheus metrics, gRPC reflection, an HTTP control plane, and a live terminal dashboard. Support is routed through GitHub Discussions and Issues, with documentation also pointing to email or a GitHub issue. This combination gives technical teams operational visibility and community channels without an indication of a dedicated support entitlement in the self-hosted plan.
The CLI runs on Linux and macOS for amd64 and arm64. The stated deployment options are self-hosted and BYOC: in the latter, EZ-CDC says the deployment runs in the customer's AWS or GCP account and data stays within that account. BYOC features listed by the maker include multi-AZ high availability, autoscaling, load balancing, optional PrivateLink, and data sovereignty. These are useful considerations for organizations with infrastructure or data-location requirements, while support and optional PrivateLink connectivity are charged separately.
Pricing
Self-host dbmazz costs nothing, billed Free forever. It includes the open-source engine, CLI and terminal dashboard, local web UI, and community support. This is the clearest fit for teams comfortable operating the software themselves; the tradeoff is that they own deployment and operations.
BYOC costs nothing, billed Free, for deployment into an AWS or GCP account. Support and optional PrivateLink connectivity are charged separately, so organizations needing those services should factor in additional costs. The Elastic License 2.0 permits commercial and non-commercial use but restricts offering dbmazz to third parties as a managed service, a material constraint for service providers.
Platforms
dbmazz is self-hosted, with Linux and macOS CLI support on amd64 and arm64. Its BYOC option runs in a customer's AWS or GCP account. The project also provides an official Docker image for running the daemon.
Who it's for
dbmazz makes sense for teams seeking log-based PostgreSQL CDC, a defined set of stable sinks, and a way to avoid operating a separate streaming platform. It is less suitable when one daemon must manage several independent jobs, when S3/Iceberg is a required destination today, or when a business intends to offer the software as a managed third-party service.
Pros and cons
- Pro: Free self-hosted and BYOC plans lower the entry cost for trying or deploying CDC, with BYOC keeping data in the customer's cloud account.
- Pro: Automatic creation of publications, replication slots, sink tables, and audit columns can cut setup work.
- Pro: Initial snapshots, schema evolution, and several stable sink options cover useful parts of a replication workflow.
- Con: One replication job per instance means multi-pipeline deployments require additional instance planning.
- Con: S3/Iceberg is only on the roadmap, limiting teams that need it as a current stable sink.
- Con: Elastic License 2.0 restricts managed-service resale, and BYOC support and optional PrivateLink cost extra.
Alternatives
Change Data Capture Software is the broader category for comparing CDC options. Choose Canal if Apache-2.0 licensing is preferable for a free, self-hosted project. Estuary is worth considering for a web or API-oriented option with a free plan and a listed $100 a month pay-as-you-go plan.
Debezium is another free, open-source CDC platform, with Apache License 2.0 and Linux, Windows, and self-hosted platform support. Decodable may suit teams seeking a managed-style web/API option: its free plan includes 24 hours or 10GiB stream retention, 20 streams, and four running tasks, while its Enterprise plan is $0.10.
LinkedIn Databus is a free Apache 2.0 open-source alternative. PeerDB offers a free open-source tier aimed at individuals, with Docker setup and CDC, streaming query, and query-layer features; its Micro plan is listed at 250. pgstream is a free Linux, macOS, and self-hosted option packaged as an Apache 2.0 CDC CLI and Go library.
Verdict
Choose dbmazz if your team wants a free, self-hosted route to PostgreSQL logical-replication CDC, its stable sinks fit your destination needs, and one job per instance matches your deployment plan. Automatic setup, snapshot support, and BYOC data locality strengthen the case. Look elsewhere if you need S3/Iceberg now, several jobs from one instance, or permission to offer the software as a managed service.