MongoDB 9.0: This Is What a Database Should Be

For years, organizations have been told to accept a series of tradeoffs when it comes to their technology stacks:

• If you want more performance, add more infrastructure.

• If you want stronger security, prepare for application constraints.

• If you want observability, you’ll have to connect more tools and translate more signals.

• If you want resilience, hire operators.

• If you want AI, rethink your entire architecture.

We disagree with these and with the assumption that as applications grow more sophisticated, database operations are doomed to become impossibly complex. The database does not have to be a bottleneck. It can and should provide the performance, control, and flexibility to move faster.

Every application, transaction, customer interaction, AI model, and agent—the database should be the engine powering these, not what’s holding them back. That's the idea behind MongoDB 9.0, the most advanced release in MongoDB's history.

The world is changing. Databases need to keep pace.

Modern applications must serve millions of users while simultaneously supporting growing fleets of AI agents. AI systems need access to fresh operational data, and agents execute actions continuously, introducing new levels of concurrency and unpredictability. Development teams are expected to move faster, while security and compliance requirements continue to rise. Yet many organizations are still stitching together separate systems for operational data, analytics, information retrieval, observability, security, and AI.

Every additional system creates complexity and an additional surface of risk. Every integration creates friction. Every bottleneck slows innovation. Organizations need a data foundation that supports applications, builders, and AI systems without requiring a new architecture for each emerging requirement in a fast-moving market.

MongoDB 9.0 is built for that reality. It advances MongoDB’s core engine across performance, scale, security, resilience, observability, and developer productivity. It is not a collection of disconnected features. It is a single, coherent experience for building and operating modern applications across MongoDB’s entire offerings. Let’s explore the key pillars that make MongoDB 9.0 the foundation for organizations pushing innovation forward instead of working around the constraints of the past.

More room to build

Performance matters when it translates into faster user experiences and business growth.

MongoDB 9.0 raises the ceiling for demanding enterprise and AI-powered applications. Compared to 8.0, transactional workloads achieve up to 20% higher throughput, enabling applications to process more business-critical operations more efficiently. Core operations are faster across the board, with find-one queries up to 35% faster and update-one queries up to 30% faster, helping applications respond more quickly to users and downstream systems. For large deployments, MongoDB 9.0 delivers nearly 2x higher throughput, enabling teams to support more users, transactions, and AI-driven activity before adding infrastructure or re-architecting applications.

For MongoDB Atlas customers, Gen2 compute is now available for M30+ dedicated clusters on Amazon Web Services (AWS) and Google Cloud Platform (GCP). Gen2 delivers improved price-performance along with more flexible storage and IOPS controls, helping organizations optimize resources for their specific workload requirements.

The outcome is simple: faster application performance, greater operational efficiency, and more room to grow within the infrastructure you already run. MongoDB 9.0 helps organizations scale further before performance becomes a constraint.

Our goal remains the same across the board: superior performance at optimal cost.

Security without compromise

As organizations collect more sensitive data, security requirements continue to grow. Too often, stronger protection comes at the expense of application functionality. MongoDB 9.0 challenges that tradeoff.

MongoDB’s industry-first Queryable Encryption now supports prefix, suffix, and substring queries, allowing applications to perform partial-match searches while data remains encrypted throughout its lifecycle. Real applications don’t search only for exact matches; they search names, identifiers, domains, and other fields the way users actually interact with them. With MongoDB 9.0, organizations can build rich search experiences over sensitive data without sacrificing security.

MongoDB Atlas also expands secure connectivity and encryption capabilities. New support for Google Private Service Connect (PSC) port mapping simplifies private networking by allowing multiple database nodes to be accessed through a single endpoint. AWS PrivateLink cross-region connectivity enables private communication between applications and Atlas clusters across AWS Regions without exposing traffic to the public internet. Azure Key Vault (AKV) encryption at rest now supports secretless authentication using an Atlas-managed Service Principal, eliminating the need to store long-lived credentials in Atlas. Atlas also adds support for Google Trust Services (GTS) as a certificate authority, providing additional resilience for TLS certificate management.

Together, these capabilities make it easier to secure sensitive data, simplify network architectures, and strengthen encryption workflows. The result is enterprise-grade security that supports modern application requirements rather than limiting them.

When pressure arrives, the database should know what matters

Production environments rarely fail because everything goes according to plan. They fail when traffic spikes, connections surge, resources contend, or workload patterns change faster than teams can respond. MongoDB 9.0 helps make those moments uneventful.

Intelligent Workload Management (IWM) is a managed availability capability that keeps Atlas clusters responsive and predictable even under overload. When a cluster is overloaded, IWM queues transient traffic spikes, keeps short-running operations moving and long-running operations stable, and throttles excess requests only as a last resort to maintain cluster stability. It buys crucial time for auto-scaling to add capacity or for an operator to act. This means that, even under overload, your application stays online and your customers never notice the difference.

MongoDB 9.0 also strengthens the controls that help teams prepare for and recover from the unexpected. For self-managed MongoDB, Enterprise Advanced now writes S3 backups over a direct data path that bypasses the Ops Manager control plane entirely for faster performance. It also produces fully immutable, portable point-in-time restore (PITR) chains in S3. Safe-sequence maintenance controls help validate infrastructure updates in non-production environments before they reach production.

Database and collection-level restores for Atlas, enhanced snapshot management, and customizable maintenance waves give teams greater control over data recovery and maintenance operations. Together, these capabilities help organizations strengthen resilience, streamline change management, and manage updates on their own terms.

Resilience becomes less of a promise and more of an operating capability: protect critical work, recover what matters, and make maintenance more controlled.

See the whole system

Modern systems generate enormous amounts of telemetry. The challenge is not collecting data. It is understanding what that data means in the context of the application. MongoDB 9.0 brings database behavior and application behavior into clearer view. A redesigned Atlas Metrics UI gives teams a more intuitive way to understand cluster health and workload performance, helping them detect and diagnose issues faster.

Enhanced native OpenTelemetry integration enables Atlas metrics, logs, and events to stream directly to any OpenTelemetry-supported observability platform, including popular options such as Datadog, New Relic, Dynatrace, Prometheus/Grafana, Google Cloud, Azure, and AWS CloudWatch.

MongoDB 9.0 also expands Query Shape metrics beyond reads to include inserts, updates, and deletes. Teams can see a more complete picture of workload performance, resource consumption, and database activity across the full application lifecycle.

When something goes wrong, organizations should not have to wonder where to look first or translate their database signals before they can act.

Modernize without starting over

Modernization should not require organizations to discard the investments that already make their teams productive. MongoDB 9.0 lowers the cost of moving forward. The MongoDB extension for Hibernate ORM enables Java developers to continue using familiar Hibernate APIs and Hibernate Query Language while working with MongoDB.

New schema constraints make it easier to enforce application expectations directly in the database, providing teams with a shared, authoritative place for data rules.

Together, these capabilities create a more practical path to modernization: preserve familiar development patterns, reduce distributed validation, and gain the flexibility of MongoDB without rebuilding application logic from scratch.

One foundation for what comes next

The next generation of applications will combine operational workloads, search, analytics, and AI. Agents will continuously read and write data. Applications will need to respond faster, operate more reliably, and protect more sensitive information, all at the same time. That future does not need more fragmentation. It needs a stronger foundation.

MongoDB 9.0 gives organizations greater capacity to grow, more control as conditions change, greater visibility into production behavior, stronger protection for sensitive data, and a lower-friction path to modernization. It strengthens MongoDB as the foundation for applications, builders, and AI systems across Atlas, Enterprise Advanced, Community Edition, and self-managed environments, including regulated and air-gapped deployments.

Because the database should never be the bottleneck.

It should be the platform that makes everything else possible.

This is what a database should be.

Note: Performance results are based on internal MongoDB testing comparing MongoDB 9.0 and MongoDB 8.0 using public benchmark methodologies. Results may vary by workload, deployment, hardware, and configuration.