Closed kalantar closed 3 months ago
@amye we believe this should be TAG App Delivery - could you please re-label?
@amye we believe this should be TAG App Delivery - could you please re-label?
@amye (or someone else), could this be done?
Adding both, trying to get more TAGs engaged in process
@kalantar Please confirm Iter8 will be available to present on April 17 in TAG App Delivery? (thank you for understanding the move from last meeting)
@kalantar Please confirm Iter8 will be available to present on April 17 in TAG App Delivery? (thank you for understanding the move from last meeting)
@angellk Unfortunately, I find that I will not be able to attend the meeting on April 17. I've reschedule for mid May instead if that is ok.
What was the result of the TAG meeting? Please add notes.
TAG-CS review, the project has:
We would like the project to reapply in 6 months and complete the following:
Application contact emails
kalantar@us.ibm.com, alan.cha1@ibm.com, spartha@us.ibm.com
Project Summary
Iter8 makes it easy to ensure that Kubenetes applications and ML models perform well and maximize business value.
Project Description
Iter8 is the Kubernetes release optimizer built for DevOps, MLOps and data science teams. It makes it easy to ensure that Kubernetes applications and ML models perform well and maximize business value. In particular, Iter8 provides mechanisms that simplify all types of application and model testing -- blue-green, canary, A/B/n and performance. It automatically and dynamically reconfigures routing resources to support application transparent blue-green and canary testing as part of application/model release. It provides a client SDK to support A/B testing of backend components. Finally, it provides composable tasks to enable rapid creation of performance tests. These techniques are integrated into Iter8.
Org repo URL (provide if all repos under the org are in scope of the application)
https://github.com/iter8-tools
Project repo URL in scope of application
https://github.com/iter8-tools/iter8
Additional repos in scope of the application
Documentation
Website URL
https://iter8.tools
Roadmap
https://iter8.tools/1.1/roadmap/
Roadmap context
No response
Contributing Guide
https://github.com/iter8-tools/iter8/blob/master/CONTRIBUTING.md
Code of Conduct (CoC)
https://github.com/iter8-tools/iter8/blob/master/CODE_OF_CONDUCT.md
Adopters
https://github.com/iter8-tools/iter8/blob/master/ADOPTERS.md
Contributing or Sponsoring Org
https://www.ibm.com
Maintainers file
https://github.com/iter8-tools/iter8/blob/master/MAINTAINERS.md
IP Policy
Trademark and accounts
Why CNCF?
Iter8 seeks a vendor-neutral home that will offer the opportunity for greater project awareness, adoption, and contributions. Iter8 will also benefit from CNCF's guidance on best practices and governance.
Benefit to the Landscape
Iter8 benefits the CNCF landscape by adding a tool that simplifies testing. Different types of testing require different approaches. Iter8 brings these approaches together in a single tool. A common testing approach simplifies adoption, ensuring user applications and models are of high quality.
Cloud Native 'Fit'
Iter8 fits in the category Continuous Integration & Delivery. Its suite of testing capabilities can be used by other CI/CD tools.
Cloud Native 'Integration'
As noted above, Iter8 complements Continuous Integration & Delivery projects. Iter8 provides a set of tools that enable various forms of testing which can used as part of CI/CD.
To implement blue-green and canary testing, Iter8 depends on a service mesh. Iter8 works natively with the Kubernetes gateway API supported by, for example, graduated projects Istio and Linkerd.
To simplify user interaction and configuration, Iter8 uses Helm.
Cloud Native Overlap
Similar projects
Flagger (sub-project of Flux) and Argo Rollouts share similarities with Iter8. Both provide support for advanced application rollout on Kubernetes with blue-green and canary analysis. They work with many service meshes and ingress products to provide this support. Users specify the desired rollout using a Kubernetes custom resource.
Iter8 was heavily inspired by both projects. However, Iter8 differs in several regards. With Iter8:
Applications can be composed of any Kubernetes resource type. For example, it works with machine learning applications built using KServe
InferenceService
resources out of the box. To do so, Iter8 allows the user to specify the resources being deployed as part of the specification of the rollout instead of assuming a particular pattern.Users can A/B/n test application backend components. Beyond providing HTTP header and cookie-based routing, Iter8 provides a client SDK with a simple API that allows users to write frontend components designed to focus A/B/n testing on the backend components.
No custom resource is required to specify rollouts. Both Flagger and Argo Rollouts require the user to install and use a custom resource definition to define rollouts. In Iter8, users simply specify rollouts using Helm configuration files.
Landscape
no
Business Product or Service to Project separation
N/A
Project presentations
Technical Introduction
Project champions
TBD
Additional information
No response