How Java Workload Optimization and OOM Response Empowers Platform Teams

We recently announced a broad expansion of our Optimize Live product, including support for Java Virtual Machine (JVM) Workload Optimization and automated out-of-memory (OOM) Response to bolster application performance, reliability, and efficiency. These enhancements further enable platform teams to dramatically cut cloud costs while improving reliability and unburdening engineers from manual toil. 

You can read the full details in the press release. This post is focused on what some of the broader trends that we've heard from platform teams about their biggest pain points and how these new features were designed to address them.

When it comes to java applications in particular, it’s really tricky to get resource management right. You need to take into account a variety of additional factors to properly optimize, especially for memory, including knowing the appropriate heap size and then sizing the container around that heap size.

It also takes a lot of courage for platform engineers to drive automated optimization across their environment when developers are often hesitant to take recommended actions and wary of potential impact on their applications.

We want our users to have confidence that our recommendations are taking their JVM resource needs into consideration and giving them automated actions that improve their reliability. StormForge’s new Java Workload Optimization brings in JVM metrics and allows holistic rightsizing of JVM workloads. Other rightsizing tools don’t dive deep enough into the JVM, so their recommendations are based solely on memory usage at the container level.

For anyone trying to rightsize JVM applications on Kubernetes, it can feel like trying to solve a Rubik’s cube with a blindfold on. That’s why we’re so proud to offer JVM Workload Optimization, giving you rightsizing recommendations that take JVM’s unique resource needs into consideration. Now you can deploy automated actions that improve your reliability with JVMs — rather than put your applications at risk of performance issues.

With any type of workloads environment, changes in memory usage occur due to changing traffic patterns over time and regular software updates. This unpredictable memory usage will eventually cause OOM kills leading to service disruptions that impact performance and business reputation.  

StormForge’s new OOM Response is like insurance for memory settings. It helps avoid application downtime by automatically detecting OOM events and increasing memory resources, ensuring stable workload operations without manual intervention. Machine learning continues analyzing the increased memory consumption to refine the recommendations over time for optimal resource allocation. Detailed reports track OOM events declining over time to measure impact as they’re automated away.

With this release, you get rightsizing recommendations with proactive optimization for all common workload types paired with reactive protection with OOM Response. This ensures that every platform team is empowered to drive automated optimization in their environment while improving reliability.

Rounding out this feature release is the addition of a Mutating Admission Webhook, which further simplifies the integration of CPU and memory recommendations with popular GitOps deployment tools, including Argo CD and Flux.