Recently, the cloud-native application management platform Rainbond officially released version 6.9.0. Through architectural restructuring and functional upgrades, it directly empowers enterprise users with the capability to deploy AI large models privately, achieving full-process management from model selection, one-click deployment, to OpenAI-compatible invocation.

For enterprises wishing to build AI capabilities within their internal resources, version 6.9.0 offers an extremely user-friendly "turnkey" solution. After enabling the built-in AI large model plugin, developers can immediately unlock four core modules in their workspace: model repository, instance management, API keys, and model monitoring. In terms of model sources, the platform supports both pre-installed common models and allows pulling models from the ModelScope community, or even deploying custom models via HTTP addresses, local paths, or direct file uploads. During deployment, the system defaults to the high-performance vLLM engine and provides flexible configuration options for both GPU and CPU environments, significantly lowering the threshold for enterprises to build private large models.

Improving management efficiency was also a key focus of this update. With the new "model instance" management panel, operations personnel can monitor instance status in real-time, view detailed logs, and quickly diagnose issues when instances encounter abnormalities. For integration with business systems, the platform provides an API interface compatible with OpenAI, allowing business parties to seamlessly connect to privately deployed models without changing their existing development workflow by simply replacing the base URL and key. The accompanying monitoring module offers comprehensive visualization from service response quality to GPU device utilization, making every bit of computing power traceable.

In addition to breakthroughs in the AI field, version 6.9.0 has also deepened its virtual machine (VM) management capabilities. To meet enterprise needs for asset migration and flexible expansion, the new version supports direct deployment of qcow2 images and introduces a hot expansion feature for VM specifications, enabling lossless upgrades to CPU and memory while the VM is running. Additionally, the introduction of multi-VM application-level orchestration and a complete set of application template delivery capabilities allows enterprises to package complex VM applications into standardized installation packages, enabling rapid recovery and migration across environments. To address the specific needs of Windows users, the platform has added the VirtIO driver disk mounting function, ensuring compatibility when deploying across different virtualization environments.

As a large-scale architectural upgrade, version 6.9.0 also fixed occasional connection issues with RainAgent at the underlying level and optimized several engineering guidance experiences, ensuring stability in production environments. Whether for private training and inference of AI models or cloud-native management of traditional VM workloads, this version demonstrates Rainbond's deep expertise in enhancing enterprise R&D efficiency. Currently, Rainbond users can experience this new version through the one-click upgrade feature on the platform management interface, embarking on a new digital transformation journey combining cloud-native and AI technologies.