Skip to content

openshift-qe/ocp-qe-perfscale-ci

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

44 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

ocp-qe-perfscale-ci

Purpose

This repo contains various Jenkinsfile and other pertaining content that are used by a multi-branch pipeline in the OCP QE Jenkins instance.

Structure

There are multiple orphan branches present in this repos, each of them are supposed to house one kind of workload for E2E-benchmarking. Each branch should contain one Jenkinsfile

You can create a new orphan branch simply by git checkout --orphan BRANCHNAME for new workload.

The OCP QE Jenkins multi-branch pipeline job will look at the Jenkinsfile on each of these branches and create a new workload job for you to execute in your Jenkins.

This repository also hosting the perf-dashboard-grafana-crs directory, that includes all the Custom Resources and relevant files that you need to deploy a fully functional perf-scale dashboard. This deployment uses grafana operator to enable grafana operator on your cluster, create an instance, create required datasources(in this case Prometheus and ElasticSearch) and dashboards.

See launch-grafana.sh for env variables that are needed. cleanup-grafana.sh can be used to clean up everything created by the launch-grafana.sh script.

PREREQUISITE: is to have KUBECONFIG env variable configured that can access your OpenShift cluster.

Jobs

This is an overview - for additional details about each job, see the respective README in the specific branch for that job. Note that some jobs may call others.

Job Purpose
benchmark-cleaner TODO
benchmark-comparison TODO
cerberus TODO
chaos-upgrade TODO
cluster-builder TODO
cluster-post-config TODO
cluster-workers-scaling TODO
etcd-perf TODO
kraken TODO
kube-burner TODO
kube-burner-ocp TODO
loaded-upgrade TODO
must-gather TODO
netobserv-perf-tests Used for Performance and Scale testing of the Network Observability Operator
network-perf TODO
network-perf-v2 TODO
post-results-to-es TODO
post-to-slack TODO
regression-test TODO
router-perf TODO
upgrade TODO
write-scale-ci-results TODO

Scripts

Mr. Sandman

Sometimes, the only way to get data such as UUID and workload timestamp information is directly from the workload job runs. If you find yourself in need of this but don't want to manually pour through logs, you can let Mr. Sandman give it a shot by running ./scripts/sandman.py --file <path/to/out/file>

By default he will output a JSON of data to ./data/workload.json but you can also output a Shell file for use in BASH environment using the --output sh flag

Authors

Kedar Kulkarni <@kedark3 on Github>

Paige Rubendall <@paigerube14 on Github>

Nathan Weinberg <@nathan-weinberg on GitHub>