January 2025 - V1.9.0

✨✨✨ The most recent release brings some solid new features, enhancements, and bug fixes that should enhance your experience. Check out what’s changed:

🚀 Highlights

Support for Alibaba Cloud ACK Optimization

CloudPilot AI now supports cost and performance optimization for Alibaba Cloud ACK clusters. In just two steps and five minutes, you can significantly reduce your ACK cluster costs while also enhancing your management efficiency.

CloudPilot AI helps you:

Scale out the most suitable ECS instances to quickly run your workloads.
Scale in nodes efficiently when utilization is low.
Monitor costs easily with a user-friendly dashboard.
Automate and manage your cluster’s data plane, including scaling and upgrades.
Customize node configurations to fit your workload needs.
Leverage Spot instances to reduce cluster costs while maintaining stability.

Enjoy a cost-efficient and effective journey on ACK with CloudPilot AI!

Proactively Maintain Workload Distribution as Configured

In earlier versions of CloudPilot AI, modifying the workload configuration (see the configuration page below), such as switching from spot-friendly to non-spot-friendly, did not take effect immediately. Instead, all changes were applied during the next rebalance operation:

Now, CloudPilot AI applies the changes immediately to ensure the workload distribution meets your requirements.

⚙️ Enhancements

Support flexible node capacity configurations, including Spot and On-Demand instances.
Reduce the delay for converting to a cheaper Spot node from 1 day to 1 minute, further optimizing overall costs.
Ignore PDB (Pod Disruption Budget) configurations. If an error occurs in the optimization plan, CloudPilot AI will provide an optimized result whenever possible.
Since StatefulSets do not support rolling updates, use the evict API directly for single-node workloads. If high availability is required, ensure that the number of replicas in the StatefulSet is greater than one.

🛠️ Bug Fixes

Ensure that Pods of the same workload (e.g., Deployment) are scheduled on at least two nodes, instead of three as in earlier versions.
Restrict subnet selection to user-configured zones to avoid node creation failures.

We believe these updates further enhance CloudPilot AI’s performance, efficiency, and usability. For any questions or support, feel free to connect with us via our Slack channel .

Stay tuned for more updates! 🚀