apache/kafka

Prefer this fork only if you specifically need its storage-tiering and multi-disk design. If you want current Apache Kafka features, bugfixes, and lower maintenance risk, upstream is the safer choice.

Prefer upstream Kafka unless you specifically need this fork's older, heavily modified baseline. Choose this fork only if you are willing to own the maintenance burden and do not need current upstream fixes.

Prefer upstream Kafka unless you specifically need this old snapshot or its local customizations; this fork is too stale for normal production adoption and is best treated as legacy code, not a current platform choice.

Choose upstream unless you specifically need the Microsoft mirror or an older divergent codebase. For most adopters, this fork is too stale and too far from current Kafka to be a safe default.

Prefer this fork if you want Kafka with extra operational automation and Confluent-oriented infrastructure. Prefer upstream if your priority is the newest canonical Apache Kafka line with the least fork-specific surface area.

Choose this fork only if you specifically want a stale Kafka snapshot or mirror. If you want an adopter-friendly Kafka distribution, current upstream is the better choice.

Prefer this fork only if ZooKeeper removal is the primary requirement and you are comfortable owning a long-lived, heavily divergent codebase. For most production adopters, upstream Kafka is the safer choice because this fork is far behind and likely missing many modern features and fixes.

Prefer upstream Apache Kafka unless you specifically need this legacy, heavily diverged snapshot. This fork is stale and materially behind current Kafka, so it is a poor choice for new deployments or active maintenance.

Choose this fork if your goal is study and annotation. Choose upstream if you need current Kafka features, fixes, and operational reliability.