[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"$f2p8xek1r3rg1g":3},{"code":4,"message":5,"data":6,"timestamp":40},10000,"success",{"full_name":7,"position":8,"description":9,"language":10,"stars":11,"forks":12,"stars_gained":13,"age_days":14,"hn_threads":15,"one_liner":16,"summary_md":17,"topic":18,"github_topics":21,"license":30,"homepage":31,"updated_at":32,"history":33},"apache\u002Figgy",14,"Apache Iggy: Hyper-Efficient Message Streaming at Laser Speed","Rust",5197,491,6,1316,[],"A Rust message streaming server for teams that want Kafka-style append-only log semantics without running a JVM stack.","Apache Iggy is worth a look if you already understand message streaming and want a Rust-native server that speaks several transports. It is not a drop-in Kafka replacement today, but it is a credible self-contained log for new pipelines.\n\n## The problem\n\nA team needs to move events from producers to several consumers with ordering and replay, for example feeding a real-time dashboard while a batch job re-reads yesterday's range. The default path is standing up Kafka or a database-backed queue, which brings broker infrastructure, JVM operations and serialization conventions. With Iggy, the broker is a single Rust binary that stores a persistent append-only log, and producers publish over TCP, QUIC, WebSocket or HTTP while consumers poll by offset, timestamp, first\u002Flast N or next N. The result is replayable streams with the same conceptual model, minus the JVM layer.\n\n## What's different\n\nThe README's central claim is that this is a purpose-built log rather than an extension over Kafka or SQL. The mechanism is a thread-per-core, shared-nothing design with `io_uring` on Linux, custom zero-copy serialization and direct binary payload handling with no enforced schema. Configurable segment size, durability and flush thresholds are set per topic, which matters when an ingestion-heavy stream and an audit stream share one server. Clustering is compiled into the same `iggy-server` binary and toggled with `cluster.enabled`, using Viewstamped Replication for quorum commits and primary election.\n\nThe trade-off is ecosystem lock-in and maturity. The Kafka wire-protocol gateway exists but the README says it is in development and not yet in a release, so existing Kafka clients cannot simply point at Iggy. Server, SDKs and CLI are versioned independently, so you must check the [server compatibility table](https:\u002F\u002Figgy.apache.org\u002Fdocs\u002Fsdk\u002Fintroduction#server-compatibility) before matching versions.\n\n## Who should try it\n\nGood fit: a team building a new event pipeline in Rust, Go, Java, Python, C# or Node.js that wants low-latency streaming and is willing to adopt a newer broker. The multi-language SDKs and the [getting started guide](https:\u002F\u002Figgy.apache.org\u002Fdocs\u002Fintroduction\u002Fquickstart\u002F) lower the on-ramp, and the bundled CLI and Web UI help with administration. Poor fit: an organization with a large existing Kafka deployment and client tooling, because the gateway is pre-release and migration is not yet a supported path. Also a poor fit for anyone who needs a schema registry enforced at the broker, since binary payloads are accepted directly.\n\n## Requirements and limitations\n\nExpect Linux to be the happy path: the README ties its best performance to `io_uring` and thread-per-core on modern Linux, so other platforms lose the headline advantage. Payloads are untyped, so schema validation and contract management are your application's job. Clustering is off by default and requires VSR configuration and quorum-aware deployment, not just flipping a flag in production. Version drift between server and SDKs is a real ongoing concern given independent semver tracks. Edge releases (`-edge.N`) track latest changes but the README explicitly warns they are not recommended for production, so pin stable releases. Connectors are Rust plugins implementing `Source` or `Sink` traits, which is flexible but adds a build step for custom pipelines. Check the [clustering docs](https:\u002F\u002Figgy.apache.org\u002Fdocs\u002Fclustering\u002Fvsr) and architecture documentation before sizing a deployment.\n\n## HotRepo's take\n\nHotRepo only has a two-day sample: 5,191 stars with +7 on 2026-10-24 and 5,192 with +6 on 2026-10-25, both on the daily board. That is insufficient comparable history to judge whether interest is sustained, so treat the trend as a snapshot rather than momentum. The specific thing to verify before adopting is the server-vs-SDK compatibility table for your chosen language; mismatched versions are the most likely first integration snag.",{"slug":19,"name":20},"data","Data",[22,23,24,25,26,27,28,29],"apache","http","iggy","messaging","quic","rust","streaming","tcp","Apache-2.0","https:\u002F\u002Figgy.apache.org","2026-10-25T03:20:43Z",[34,38],{"date":35,"stars":36,"stars_gained":37},"2026-10-24",5191,7,{"date":39,"stars":11,"stars_gained":13},"2026-10-25",1792912505]