The Data Platform Bar Has Risen. Ververica Has Set It: Introducing Ververica Platform 3.1

The Hard Part Was Never The Stream. Ververica Platform 3.1 Proves It.

In 2026, the hardest part of stream processing isn't the stream processing. It's everything other vendors don't tell you about.

It’s the surrounding ecosystem, the reliability, and the need for governance. The enterprise-grade infrastructure required to make complex systems truly trustworthy, not just to your engineers, but to the organizations that have the most to gain from real-time data.

Ververica created Apache Flink. We wrote the standard. And for over a decade, we have been solving what others couldn't: exactly-once fault tolerance, terabyte-scale state management, the edge cases that break anyone else. Then we created Ververica’s Unified Streaming Data Platform, and now, we’ve further improved the Platform with additional features.

Our expertise is not a footnote. It is a foundation, and what we’re shipping now continues to build on that same core:

A data platform where your teams move faster. Your compliance team sleeps soundly. And your business trusts every insight that comes out of the pipeline.

No hedging. No half-measures. No compromise.

This is governed, absolute truth, delivered at the speed of now.

Stop evaluating. Start running.

Image One: Ververica Platform 3.1

This Is What Enterprise-Grade Actually Looks Like

Ververica has a long history of building an enterprise-ready platform that even the most regulated enterprises rely on. The features in this new release are no exception. Every feature was built with a single question in mind: What does it take for a regulated enterprise to trust this platform in production?

The answer is the full stack of operational, security, and compliance capabilities that large organizations require before they can move fast with confidence.

Here are the new features we are delivering today:

Kubernetes-Native Operations: Stream Processing As A First-Class Citizen

Your infrastructure teams run Kubernetes. Ververica does too, natively. The Ververica Kubernetes Operator lets your teams manage the full platform directly through kubectl and Helm. No separate toolchain, and no web UI required for automation. The platform fits into the operational patterns your engineers already use. Stream processing becomes a first-class citizen in your infrastructure stack.

What this means for your business:

• CI/CD pipelines that deploy and version streaming jobs like any other application
• Automated monitoring and scaling through the tooling your team already owns
• Simpler version management and rollbacks

Governance, Traceability, and Security Built-In

Ververica’s Audit Logs and API Tokens improve governance, traceability, and secure CI/CD automation, because compliance isn't a business constraint. It's a competitive advantage that Ververica is built to address. Let’s dig in: 

Effortless Compliance with Audit Logs

Regulated industries have to address one specific question before adding any new technology: “Will we be able to audit it?”

The answer is now a resounding yes.

Ververica Audit Logs capture the events required by all of your security and compliance policies, and store them where your teams can retrieve, query, and retain them for as long as your policy requires. No workarounds. No exporting logs to a third-party system and hoping they stitch together nicely. For industries running under DORA, GDPR, SOC 2, PCI DSS, or any framework that demands evidence of data access and control, this is foundational.

Secure Machine-To-Machine Authentication

API Tokens enable secure, automated access to the Ververica Platform without requiring user interaction. The newly introduced flow for API Token Management allows you to give each workflow, integration, and CI/CD pipeline its own scoped identity. Each token is bound to a namespace and assigned a specific role (viewer, editor, or owner), and you can create, use, or revoke them at any time to comply with least-privileges policies. This makes it easy to integrate Ververica’s Unified Streaming Data Platform into automated workflows while maintaining strict access control and security.

VERA Engine: Now Compatible With Java 17 And Built For How Engineers Actually Work

The VERA engine doesn't wait. Neither do we.

By customer demand, our proprietary Platform engine VERA is now compatible with JDK 17,  the recommended runtime for Flink 2.0, because staying current isn't optional when your pipelines are mission-critical. This updated VERA version also adds native support for the bitmap type and related functions, and is purpose-built for real-time deduplication scenarios at scale.

The engine is smarter. The developer and engineering experience is sharper. The result is a data platform your teams will actually want to use.

This is what it looks like when the people who built Flink keep building.

Expanded Connectors and Catalogs: Your Data Lives Everywhere. Ververica Does Too. 

Real-time data isn't confined to a single system. Neither is Ververica.

We continue expanding the connector and catalog ecosystem with highly-requested capabilities in this latest version: 

• MongoDB Catalog
• Azure CosmosDB Connector

MongoDB Catalog

No more hand-crafting DDL for every collection. The MongoDB Catalog connects your databases directly to Flink SQL, so databases map to databases, collections map to tables, schemas infer automatically. Your data is ready to query before you've finished your coffee.

Azure CosmosDB Connector

Your Flink pipelines now write directly to Azure Cosmos DB with no middleware, no workarounds. Built on the AsyncSink framework with the Cosmos DB Java SDK v4 async client, it delivers at-least-once semantics with the throughput your pipelines demand. Point it at your NoSQL API endpoint and ship.

No lock-in. No latency. No data left behind.

A CLI That Developers Will Actually Use

Developer ease and productivity aren’t bonus features or nice-to-haves. They are where adoption lives or dies.

For that reason, we are releasing the Ververica Command Line Interface (vvctl) that gives your developers a familiar interface, similar to kubectl, for every platform resource. This makes it seamless for human developers to operate, but also opens the door to automated, real-time monitoring and operation directly from the terminal, and makes it much easier to integrate with your existing CI/CD flows.

We often see that developers don’t want to click through a UI to understand what a streaming job is doing, but rather prefer to query it, pipe it, and script it. vvctl allows you to do exactly that, fitting how modern engineering teams actually work.

Platform Operations, Tightened.

Shipping fast means nothing if you can't operate at scale. These additional features give your platform and infrastructure teams sharper visibility, safer credential handling, and more flexible state recovery. These are the controls that turn a capable platform into a production-grade one.

Blob Credentials Using Mounted Files

Hardcoded credentials in config files aren't security, they're a liability waiting to surface. Mount your storage credentials as files instead, each key stored separately and loaded automatically by the platform. Built for environments where credentials rotate frequently or live outside the cluster. Secure by design, not by hope.

Loading Savepoints From Custom Locations

State doesn't care where it lives, and now, neither does Ververica Platform. Load a savepoint from any accessible storage location, including deployments migrated from Ververica Platform 2.0, directly through the UI or Kubernetes Operator. Pick up exactly where you left off, no matter where "left off" happens to be.

Resource Usage Tracker

Running infrastructure across teams without visibility isn't operations, it's guesswork. The Resource Usage Tracker gives you per-namespace CPU core consumption over time, exportable as CSV via a single API endpoint. Chargeback, optimization, transparency: now you have the data to back every conversation.

Raising The Bar With Ververica Platform 3.1

For years, the  industry has debated capability. “Can stream processing do it? Is it fast enough? Reliable enough?” 

Ververica Platform 3.1

That conversation is over. Ververica has the answers to these important questions: 

• Can you deploy your data platform the way you deploy everything else?
• Can you audit it?
• Can you secure it?

• Can your developers work with it without learning a new paradigm?

This release says yes to it all.

We have the technology. We have the pedigree. We have the proof. 

What Deployment Is Best For You?

One solution, multiple deployments. We’re on our way to full feature parity. Here is the exact picture of the new Individual features available in Ververica’s Unified Streaming Data Platform  by deployment model: 

Feature	Ververica Platform:    Self-Managed	Ververica Cloud: BYOC	Ververica Cloud: Managed Service
Audit Logs
API Token Management
Kubernetes Operator		-	-
vvctl CLI
VERA Engine (v. 4.5) + Java 17
Mounted Blog Credentials		-	-
Resource Usage Tracking		-	-
MongoDB Catalog
CosmosDB Connector

More Resources

• [Product Webinar & Demo] Register: Dual Pipelines Are Done: Unified Streaming & Batch (Now)

• [Release Notes] Read the technical docs

• [Learn More] Check out Ververica Platform 3.X