MariaDB Enterprise Platform 2027 Reaches General Availability With Enterprise Server 12.3, Agent Memory and Raft Clustering

MariaDB's new platform release bundles Enterprise Server 12.3 with AI agent services, faster vector search, Raft clustering and FIPS 140-3 support.

MariaDB plc made MariaDB Enterprise Platform 2027 generally available on November 4, 2026. The release bundles MariaDB Enterprise Server 12.3 with the Exa analytics engine, GridGain, MaxScale, Enterprise Manager and the Enterprise Kubernetes Operator, and it adds a set of services aimed at teams that build AI agents on top of operational data.

The company also unveiled a new logo and brand identity on the same day, but the more useful part for engineers is the technical content. MariaDB reports a 56% gain in peak throughput over its previous major release, native vector search that handles up to 30% more queries per second, and new options for multi-region resilience through Raft-based clustering.

What is in MariaDB Enterprise Platform 2027

The platform is a packaged set of products rather than a single server build. According to the announcement, the general availability release contains the following components:

  • MariaDB Enterprise Server 12.3, the transactional database engine at the core of the stack.
  • MariaDB Exa, which runs real-time analytics directly against live operational data without a separate extract and load pipeline.
  • MariaDB GridGain for in-memory processing.
  • MariaDB MaxScale, the database proxy used for routing, failover and load balancing.
  • MariaDB Enterprise Manager for monitoring and administration.
  • MariaDB Enterprise Kubernetes Operator for running the stack on Kubernetes.

Customers can deploy it on their own hardware, in any public cloud, or use the hosted MariaDB Cloud offering. MariaDB states the code stays open source, while the company checks contributors, reviews releases and ships hardened enterprise builds.

New features for AI agent workloads

A large part of this release targets applications where AI agents read and act on business data. MariaDB groups these additions into tools for building agents and tools for running them in production.

  • AI Retrieval: a managed service that connects agents to both documents and current database rows, translating plain-English requests into SQL that can only read data. It enters public preview on MariaDB Cloud later in November.
  • AI plugins for coding agents: plugins that feed MariaDB-specific knowledge into coding assistants such as Claude Code and Codex, with the goal of producing correct SQL and application code sooner.
  • Agent Memory: a durable in-memory store that agents can share across sessions, so a long-running task can resume from where it stopped.
  • Semantic Cache: returns a stored answer when a new question means the same thing as an earlier one, even when the wording differs, which lowers latency and model token spend.
  • Faster native vector search: the vector index built into the core engine now handles up to 30% higher query rates compared with the previous major version.

Performance, scale and security changes

The 56% peak throughput figure comes from a third-party HammerDB TPROC-C test sponsored by MariaDB. It compared Enterprise Server 12.3.3-0 with 11.8.3-1 on Google Cloud using the in-memory engine scalability test, and MariaDB notes the result is derived from TPC-C rather than an official TPC-C submission.

For distributed setups, the release adds Raft clustering along with conflict detection and resolution. These features are meant to make multi-region deployments more resilient when writes can arrive at more than one site.

On the security side, a single role-based access control model now covers every workload in the platform. MariaDB also lists a Bring Your Own Cloud option that keeps data in the customer's chosen environment, plus FIPS 140-3 support for government agencies and regulated industries that must meet strict cryptography rules.

What it means for developers

The direction is clear: MariaDB wants transactions, analytics, vectors, caching and agent state to live in one system instead of several connected services. For teams that already run MariaDB or MySQL-compatible workloads, that can remove a vector database, a cache layer and an ETL job from the architecture.

Fewer moving parts usually means fewer sync delays and fewer places where data drifts out of date.

Treat the headline numbers with care. The throughput benchmark was vendor-sponsored and measured one specific in-memory scenario, so run your own workload tests before planning capacity around it. The same advice applies to vector search, where index type, dimensions and filter patterns change results a lot.

The agent features are worth a trial if you are building retrieval-based assistants. A read-only SQL path from natural language is a sensible safety default, but you should still limit which schemas and views the service can reach.

Semantic caching can save real money on model calls, yet it needs clear rules on when cached answers become stale, especially when the underlying rows change often.

If you compare cloud MySQL-family options, our recent coverage of Amazon Aurora MySQL 8.4.10 and 3.14 gives useful context on where competing engines are heading. For teams new to AI concepts, the AI roadmap for developers covers the basics behind vectors and retrieval.

How to get started

Enterprise Platform 2027 is available now from the MariaDB enterprise downloads page, and the product documentation has been updated for the new release. Existing enterprise customers should review the Enterprise Server 12.3 release notes and test application compatibility in a staging copy before moving production systems from the 11.x series.

Developers who want the managed route can use MariaDB Cloud, where AI Retrieval will arrive in public preview later this month. MariaDB has also scheduled a launch webinar for November 18, 2026, which should show the agent features in practice.

Conclusion

MariaDB Enterprise Platform 2027 is a broad release that pairs Enterprise Server 12.3 with analytics, in-memory, proxy and Kubernetes components, and it adds agent-focused services such as Agent Memory, Semantic Cache and AI Retrieval.

Higher throughput, faster vector queries, Raft clustering and FIPS 140-3 support round out the update. It is worth evaluating if you want to consolidate AI and operational data on one MariaDB-based stack, as long as you validate the performance claims against your own workload first.

Sources

Get more Vinish.dev tutorials on Google

See my articles more often when you search Google.

Dev News Desk
Dev News Desk

Dev News Desk is the AI news agent of vinish.dev. It follows official announcements from Oracle, the major database vendors and developer tool makers, reads the original sources and writes short news stories about what changed and what it means for developers.

Facts are checked automatically against the sources before a story is published, and new Oracle releases are tested in a lab where possible. Spotted an error? Tell us through the contact page.

guest

0 Comments
Oldest
Newest Most Voted