Automate Performance Management In KubernetesRightsize and autoscale containers with machine learning for performance and cost. Get resourcing right from the beginning.
dollar_sign Reduce cost by up to 40%
activity Increase Performance by 70%
crosshair Guarantee Service Level Objectives

Maximize Cluster Utility

What makes Valence Unique?

Declare Performance

Operators use Valence by declaring the performance they want to see from their applications. It's Valence's job to continually learn the optimal requests, limits, and replicas for that declared performance.

Scale Applications Exactly

Valence models how applications behave and sets resources accordingly. Valence learns how applications behave according to their workload and finds the most appropriate autoscaling configuration, both vertically and horizontally.

Optimize Resourcing Continually

Application behavior and workloads are constantly changing. Valence models application behaviour on the fly and acts as an autoscaler and rightsizer to make sure applications are scaled correctly for current and upcoming workloads, without manual intervention.

Improve Performance and Cost Effeciency

By resourcing applications exactly for their current and upcoming workload and performance objectives, Valence increases cluster utility. Valence resources applications such that they are neither over-provisioned, under-provisioned, nor causing unusable capacity on nodes.

Try Valence with your cluster to see real benefits

Get Started

How Valence works

Valence learns application behavior dynamically and operates autoscaling and rightsizing according to how your application behaves, not according to arbitrary threshold or utilization metrics. Valence continually learns better autoscaling procedures so you aren't stuck updating configuration everytime your code or workload changes.Learn More



Right Size, Optimize and Autoscale
Unlimited Deployments
Grafana dashboards for visualizing ROI with Valence
Start for free


Integration and Scalability Support24/7 supportAccount ManagementCustom Reporting, SLIs, Metrics, and SLOs
Contact Us

Frequently asked questions

Contact us for more information

Does Valence work for stateful applications?

Not currently. We focus on stateless HTTP applications but are developing the technology to work for stateful applications as well.

What if I don't have a Service Level Objectives (SLOs) for my application yet?

We don't expect you to! You can still use Valence without having "formal" SLOs defined - think of our SLO manifests as a chance to declare performance, just like you'd declare any other resource in Kubernetes. You will have at least some idea of how your application should perform - start there. A good rule of thumb is the oft-quoted human-computer interaction figure: humans notice more than 100ms of latency.

Why shouldn't I just use the Vertical Pod Autoscaler (VPA) and Horizontal Pod Autoscaler (HPA)?

Valence isn't meant as a replacement of the VPA or HPA but a mechanism to keep those core components configured and up to date with the optimal configuration for your application. You could just use the VPA and HPA together instead, but you'd be continually adjusting and monitoring them as your code and workload changes.

Valence Logo