Viduli August Update: Kafka, Project Templates, and Metrics

After many months of building, we are happy to announce major updates to the platform.

This August update brings three major improvements: managed Kafka for event-driven workloads, project templates for faster onboarding, and Prometheus-backed Metrics for richer observability. Each stands on its own, and together they make Viduli a more complete place to build and run software.

What’s new

Managed Kafka is now live (Surge)

Surge now supports Apache Kafka, with managed Kafka clusters available directly inside your project. After about three months of closed testing, Kafka is now generally available.

With managed Kafka, you can create a cluster from the console by choosing a layout, sizing your nodes, and setting topic defaults. Viduli handles the provisioning, lifecycle, and the day-to-day cluster operations, so you can spend more time on the events that define your product.

Surge runs on Kafka 4+ with KRaft, so you get modern Kafka without ZooKeeper overhead.

You can also pick a topology that fits your workload:

• Combined nodes for controllers and brokers on the same pool, great for testing and lighter workloads
• Independent pools with separate controller and broker nodes when you need clearer scaling and production-style operation

Kafka topic management is also built into the console. Open the Topics tab on your Kafka resource to create, update, and delete topics. Configure defaults like partitions, replication settings, cleanup policy, and other topic settings, then apply them consistently as your streams grow.

Project templates are live

Starting a new project should feel fast, not procedural. With project templates, you can create from a starter instead of a blank project.

During Create Project, choose Project Template and pick a starter from the gallery. Connect GitHub, and Viduli clones the full stack into your account: apps, database, secrets, canvas layout, and private repos. Then Viduli provisions everything for you.

First launch typically takes about 5-10 minutes while resources come online.

For now, the first template is the AI Chatbot - React, FastAPI starter with a Postgres database, a solid path from zero to a running AI app on Viduli.

Metrics is now Prometheus-backed

Resource performance views got a clearer name and a stronger foundation. The Monitoring tab is now Metrics, and charts are powered by Prometheus. That means fresher time-series data and signals that go beyond basic CPU and memory.

Open any resource, switch to Metrics, and pick a time range to explore. Across the platform, you can see:

• Apps: requests and latency, including tail signals like P99 latency
• Postgres: connection and cache behavior, plus transactions and database size
• Redis: hit rates and commands/second
• Kafka: broker throughput and under-replicated partitions

And to make debugging feel less like hunting through filters, metrics are now easy to reach from where you’re already looking.

If something looks off on one instance, the Instances table includes quick links for Metrics, Logs, and Shell. Click through and you land on that view with the instance already selected.

Additional improvements

Beyond the headline changes, a few more improvements landed across the platform.

Migrated to Google Cloud. Viduli now runs on Google Cloud, giving the platform a stronger foundation for reliability and smoother day-to-day operations behind the scenes.

Updated default runtimes across Ignite. Default language and runtime images were bumped across supported frameworks, so new deploys land on current language releases without extra configuration.

Stability and console UX. A round of stability fixes and web console UX improvements shipped across the platform, making everyday workflows a little smoother.