The open source project that started it
If you search for open source Kafka backup on GitHub, you land on the community kafka-backup project by Anatoly Zelenin and Sönke Liebau. It deserves real credit: it was among the first tools to back up consumer group offsets alongside topic data, and it deployed simply as two Kafka Connect connectors. But its author has stepped back from the project. Here is what it did well — and what a maintained tool changes.
Where the community project came from
The community kafka-backup project is built as two Kafka Connect connectors: a sink connector that backs topic data up, and a source connector that restores it. It was one of the first open source tools to preserve consumer group offsets, not just records — a genuinely hard problem it solved years before most alternatives. For a small cluster and a spare Connect worker, it was a pragmatic, free way to get topics onto disk.
Its author has since announced he is no longer maintaining the project and points people to a commercial successor. That is not a knock on the work; it is the normal life cycle of a single-maintainer open source tool. But it changes the calculus for anyone choosing a backup tool for production in 2026: an unmaintained Connect plugin inherits every Kafka Connect version constraint and gets no fixes for new broker behaviour.
OSO Kafka Backup is a maintained, standalone tool written in Rust. It backs up topic records, consumer group offsets, and topic configuration to object storage (S3, Azure Blob, GCS) or local disk, and restores any of it — to the same cluster or a different one — with millisecond precision. It needs no Kafka Connect runtime. See the disaster recovery use cases for the failure scenarios it covers.
| Feature | OSO Kafka Backup | community kafka-backup |
|---|---|---|
| Backs up topic data | Yes | Yes |
| Consumer group offsets | Built-in | Built-in (Connect 2.4+) |
| Point-in-time recovery | Yes | No |
| Runtime dependency | None (standalone) | Kafka Connect worker |
| Storage backends | S3, Azure, GCS, local | Filesystem via Connect |
| Compression | Zstd, LZ4 | Connect converter |
| Cross-cluster restore + remap | Yes | Limited |
| Actively maintained | Yes | No |
| Licensing | Open source (MIT) | Open source (Apache 2.0) |
When to choose which
Choose OSO Kafka Backup when you need a maintained tool, point-in-time recovery, backups to object storage rather than a Connect filesystem sink, or a workflow that does not depend on a Kafka Connect runtime you also have to operate.
Consider the community project when you are studying how offset-aware backup works, or you have an existing deployment on a pinned Kafka version and accept that no further fixes are coming. Its documentation is still a good read on the offset problem.
Both are open source, and both back up offsets. The difference that matters for production is maintenance and independence from Kafka Connect — see the full alternatives comparison for how Connect sinks and replication tools fit alongside a dedicated backup tool.
Frequently asked questions
Is the community kafka-backup project still maintained?
No. The author has publicly stated he is no longer maintaining the kafka-backup project and now recommends a commercial alternative. The repository and its documentation remain online and are still useful as a reference, but it receives no further fixes or releases.
What did the community kafka-backup project do well?
It was among the first open source tools to back up consumer group offsets alongside topic data, and it deployed simply as a pair of Kafka Connect connectors — a sink for backup and a source for restore. For small clusters it was a practical, free option.
Does OSO Kafka Backup require Kafka Connect like the community tool does?
No. The community project runs as connectors inside a Kafka Connect worker, so you operate and version that runtime too. OSO Kafka Backup is a standalone Rust binary that connects to any Kafka cluster as a client, which removes the Connect dependency entirely.
Can OSO Kafka Backup restore to a precise point in time?
Yes. Restores can stop at a precise millisecond, so you can roll a topic back to the moment before a bad deploy or an accidental delete. The Connect-based community tool restores from a backup but was not designed for millisecond point-in-time recovery.
Ready to protect your Kafka data?
Take your first backup in minutes with the open source CLI, or talk to us about Enterprise features like encryption, RBAC, and audit logging.