Tools, Ansible playbooks, and scripts to prepare, load, and analyze a large-scale OCP cluster with ACM (Advanced Cluster Management) for cluster deployments, upgrades, and policy workloads. Supports deploying SNO (Single Node OpenShift), Compact, and Standard (MNO) clusters via Assisted Installer (AI) or Image-Based Installer (IBI), using direct manifest apply, ClusterInstance API, or ArgoCD/ZTP GitOps workflows.

Table of Contents

• Workflow Overview
• Prerequisites
• Setup
• Configuration Files
• Documentation
• Workload Scripts
• Manifest Generation Scripts
• Analysis Scripts
• Graphing Scripts
• Other Scripts
• Patch Scripts

Every scale test generally follows three phases:

Phase 1: Setup            Phase 2: Run               Phase 3: Analyze
──────────────────        ──────────────────         ─────────────────────
Generate manifests  →     Deploy clusters at    →    Prometheus metrics
Setup ZTP/ArgoCD          scale via workload         Cluster install stats
Deploy ACM hub            scripts                    Timing analysis
Prep Seed Image                                      Graphs and reports

1. Setup — Use Ansible playbooks to generate deployment manifests, configure ZTP, and deploy ACM/MCE on the hub cluster
2. Run — Execute Python workload scripts to deploy managed clusters in batches at configurable intervals
3. Analyze — Run analysis scripts to query Prometheus, compute timing statistics, and generate graphs

• Jetlag has provisioned your infrastructure — OCP hub cluster installed, bastion accessible, hypervisor nodes configured. See jetlag for provisioning guides.
• oc CLI installed and authenticated against the hub cluster
• Python 3 on the bastion machine
• Chrome (for static plotly graph export via kaleido)

Bootstrap the Python virtual environment and install dependencies:

[root@<bastion> acm-deploy-load]# ./bootstrap.sh
[root@<bastion> acm-deploy-load]# source .venv/bin/activate
(.venv) [root@<bastion> acm-deploy-load]#

Then follow the quickstart guides in Documentation to set up and run your test.

Deploys SNO, Compact, or Standard clusters via Assisted Installer or Image-Based Installer in configurable batches at timed intervals.

Deployment Methods:

Workload Phases:

1. Deploy Phase — Apply manifests or push to GitOps to deploy clusters
2. Wait for Cluster Install Completion
3. Wait for DU Profile Completion (optional)
4. Report Card / Graphing

Deploys Telco Core MNO clusters via GitOps in configurable batches, and runs a policy churn workload that periodically updates a ConfigMap triggering policy re-enforcement across managed clusters. Used for capacity guidelines testing with two test profiles:

Three Phase Run — Full capacity guidelines test with certificate rotation measurement:

1. Cert Rotation / Quiesce — 25-hour quiet period to capture certificate rotation resource spikes as a baseline before deployment begins
2. Deployment Batches and Policy Churn — Deploy 1 batch per 12-hour period with concurrent policy ConfigMap updates. Prometheus measurements are taken between phase and batch boundaries to determine resource requirements as cluster count and workload increases.
3. Steady State Measurement — Post-deployment measurement period with all clusters managed and policy churn continuing

Two Phase Accelerated Run — Faster test that completes before certificate rotation:

1. Deployment Batches and Policy Churn — Deploy 1 batch per 10-minute period with concurrent policy churn. After the last batch deploys, policy churn continues for 6 hours while clusters finish deploying and applying policies.
2. Steady State Measurement — Resource consumption levels off as policy churn ends and the hub cluster reaches steady state

Imports (manages) previously deployed clusters into ACM in batches on an interval while running a concurrent policy churn workload. Updates a ConfigMap on interval, causing policies to cycle between compliant and non-compliant states across the managed clusters.

See each script's --help output for detailed usage.

Version-specific scripts in the patch/ directory provide memory limits tuning or image patches generally used with experimental builds of specific versions to test fixes during scale testing.