Optimize Pro Architecture

StormForge Optimize Pro is composed of three parts:

  1. The StormForge Optimize Pro controller, which runs on your cluster
  2. Two Kubernetes custom resource types, called Experiments and Trials
  3. The StormForge cloud API, which uses machine learning to guide the experiment process

Additionally, there is a stormforge CLI tool for installing and interacting with the controller, as well as with the cloud API.

The StormForge Optimize Pro controller is a Kubernetes operator written in Go. The controller Pod runs in a dedicated namespace stormforge-system, and uses the Experiment and Trial resources. Users create Experiments using kubectl, then the controller automatically creates a series of Trials as it empirically searches for and finds optimal configuration values for the Experiment.

The StormForge machine learning algorithms are used to generate the specific parameter values for each new trial, as Optimize Pro searches the experimental problem space.



StormForge
Cloud API
StormForge...
Trial 1
(complete)
Trial...
Trial n – 1
(complete)
Trial...
the controller may patch
other Kubernetes resources
as as part of running a trial
the controller may patch...
Trial n (in progress)
Trial n (in progre...
User
User
stormforge-system
stormforge-system
trial jobembeddedprometheus(optional)
Kubernetes cluster
Kubernetes cluster
Optimize Pro controllerExperiment
Text is not SVG - cannot display
Last modified November 1, 2022