Hey Ampere Developers,
To commemorate KubeCon EU 2026, happening this week, we dedicate this issue to how to efficiently deploy your applications to Kubernetes. Whether it’s tuning your JVM and Kubernetes pod configuration to ensure that you are efficiently using the resources that you are paying for or getting your applications running on heterogeneous Kubernetes clusters with Arm64 and x86 compute nodes, you’ll find something useful here.
Deploying Java applications on Arm64 with Kubernetes
Hot off the presses this week, we have an in-depth guide to matching application, Kubernetes, and host compute configuration to optimize hardware usage and performance for JVM applications: https://amperecomputing.com/tuning-guides/deploying-java-applications-on-arm64-with-kubernetes
A recent study by Akamas and Microsoft found that over 50% of all Java applications running on Kubernetes have a resource limit of less than 2 CPUs and a heap allocation of less than 1 gigabyte. The same report compared a reference Java application running in two pods, each with 3 CPUs allocated, to the same application running across six pods with a resource limit of only 1 CPU. They found that this can result in a massive increase in tail latency and a decrease in throughput. What this means in practice is that most Java applications in the cloud are running inefficiently – using more resources than they need to and running slower.
This in-depth tuning guide shares important considerations when deploying applications to Kubernetes incorporating both efficient use of resources and application performance, including:
- Container awareness in the JVM, and using CPU and memory quotas and JVM configuration to ensure that your application is well behaved
- Using the CPU Manager and CPU pinning to dedicate cores to your Java applications
- Workload placement in Kubernetes, matching applications with the compute resources they need
- Configuring and tuning compute nodes with tuned
Migrating your application to Ampere instances on OCI with Kubernetes
For a practical example of how to take advantage of high performance and terrific price-performance, this self-paced tutorial shows how to use Kubernetes deployment and workload placement policies to live-migrate a WordPress database from x86 to Arm64 instances on OCI with zero downtime. Find out for yourself why so many people are adopting Arm64 instances for their cloud native workloads.
Continuous Delivery on Multi-architecture Kubernetes with ArgoCD
If you prefer video tutorials, you might enjoy this presentation from Dave Neary at Configuration Management Camp that talks through all of the steps needed to migrate container builds, CI, and CD pipelines to add Arm64 as a deployment target, and how using ArgoCD can make the process easier.
Your Voice Matters: Join the Ampere Developer Community!
Got questions? Want to share your insights? Or just looking to connect with fellow Ampere enthusiasts and experts? Our Ampere Developer Community Forum is the place to be!
Join the conversation: https://community.amperecomputing.com
Happy coding and see you in the next edition!
The Ampere Developer Relations Team