Performance enhancements for OpenShift Pipelines

Goal:
Determine performance and scale of Tekton on Openshift. OpenShift Pipelines is the productizes Tekton on OpenShift as the preferred CI/CD solution on OpenShift and is moved to Tech Preview on OCP 4.4.

The targeted data plane metrics are:

• Number of concurrent PipelineRuns
• Number of concurrent TaskRuns
• Latency for PipelineRuns to start as number of PipelineRuns increases
• Latency for TaskRuns to start as number of TaskRuns increases
• Duration of PipelineRun execution as number of PipelineRuns increases

The targeted control plane metrics are:

• Maximum namespaces per cluster that can have a Pipeline and PipelineRun
• Maximum number of resources per cluster
  • Pipeline
  • PipelineRun
  • Task
  • TaskRun
• Max number of PipelineRuns for a single Pipeline
• Max number of TaskRuns for a single Task

Problem:
Need to establish the scale of Tekton on OpenShift

Why is this important: 
In order for the pipelines team to have a baseline of performance for Tekton and address issues given that most customers will run Jenkins and other CI systems in parallel and would expect similar or better scale with Tekton in order to adopt it.

Dependencies (internal and external): 

Prioritized epics + deliverables (in scope / not in scope):

• A tool for generating Tekton load

Open questions: