Redis 8 GA — Redis becomes a queryable, vector‑ready core database with major performance gains

Key update

Redis 8 is now generally available as Redis Open Source. The release folds a Redis Query Engine and multiple formerly‑external data structures into the core distribution (JSON, time‑series, probabilistic structures, and a preview vector set), adds eight new data types, and delivers "more than 30" performance and resource‑use improvements (big wins in query throughput, replication, and command latency). The project also formalizes a name/license change: Community Edition is now Redis Open Source and the distribution lists RSALv2 / SSPLv1 / AGPLv3 options for binaries. (redis.io)

Why it matters

This is a practical inflection point for teams building search, analytics, and AI‑adjacent services. Instead of stitching Redis + separate vector DB or search service, you can evaluate a single Redis 8 deployment that natively supports queryable JSON, time‑series, approximate data structures (Bloom/Cuckoo/etc.), and an integrated query engine tuned for both horizontal and vertical scaling. That reduces architectural complexity, lowers cross‑system consistency headaches, and can materially cut latency and operational overhead — but it also changes upgrade and packaging considerations: the new redis‑full binary and redis‑full.conf (and new ACL categories) mean you should test compatibility with existing clients, modules, and CI/CD packaging. The license options matter for redistribution and downstream packaging decisions; audit your legal/packaging requirements before swapping binaries in production. Operational advice: validate performance and memory tradeoffs in staging, confirm backup/replication behavior against your RPO/RTO targets, and review ACLs (new @search/@json/@timeseries categories) before rollout. (redis.io)

Source

• Source