Scylla Cloud Automation

This project automates the deployment and benchmarking of ScyllaDB clusters on Scylla Cloud, along with provisioning of loader infrastructure and complete testing workflows.

Features

• Provision Scylla Cloud clusters using Terraform
• Automatically configure loader infrastructure for stress testing
• Full Ansible integration for provisioning and testing
• Cluster lifecycle operations (setup, scale, destroy)
• Multi-cluster management via a unified variables.yml
• CLI and web UI support

Prerequisites

• Linux (Ubuntu/Rocky/Fedora/MacOS/WSL2)
• Python 3.12.x (3.13+ not supported)
• pip3
• Terraform
• Ansible
• AWS CLI (configured)
• Scylla Cloud API Token

---

Setup

1. Clone the Repository

git clone ...

2. Python Environment

OS-specific dependencies

Show install commands for your OS

Ubuntu/Debian/Mint

sudo apt update
sudo apt install -y make build-essential libssl-dev zlib1g-dev \
libbz2-dev libreadline-dev libsqlite3-dev wget curl llvm \
libncursesw5-dev xz-utils tk-dev libxml2-dev libxmlsec1-dev \
libffi-dev liblzma-dev git

Fedora

sudo dnf install -y gcc gcc-c++ make zlib-devel bzip2 bzip2-devel \
readline-devel sqlite sqlite-devel openssl-devel xz xz-devel \
libffi-devel findutils tk-devel libuuid-devel libnsl2 git

Rocky/CentOS/RHEL

sudo dnf install -y epel-release
sudo dnf install -y gcc gcc-c++ make zlib-devel bzip2 bzip2-devel \
readline-devel sqlite sqlite-devel openssl-devel xz xz-devel \
libffi-devel findutils tk-devel libuuid-devel libnsl2 git

Suse

sudo zypper install -y gcc make zlib-devel bzip2 readline-devel \
sqlite3-devel openssl-devel xz xz-devel libffi-devel libnsl-devel \
libuuid-devel libtirpc-devel tk-devel wget git curl

MacOS

/bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/HEAD/install.sh)"
brew install openssl readline sqlite3 xz zlib bzip2 git

All Platforms (Common)

Install pyenv and pyenv-virtualenv:

curl https://pyenv.run | bash

Add the following to your shell configuration (~/.bashrc, ~/.zshrc, etc.):

export PATH="$HOME/.pyenv/bin:$PATH"
eval "$(pyenv init --path)"
eval "$(pyenv init -)"
eval "$(pyenv virtualenv-init -)"

Restart your shell or source your shell config:

source ~/.bashrc  # or source ~/.zshrc

Install Python 3.12.x and create a virtual environment:

pyenv install 3.12.10
pyenv virtualenv 3.12.10 sc-env

Note: The name sc-env is just a matter of preference, you could name it anything such as pyvenv312, if you want to.

Activate the virtual environment:

pyenv activate sc-env

Ensure the correct Python version is active:

python --version  # Should display Python 3.12.xx

Install project dependencies:

pip install -r requirements.txt

You are now ready to use the CLI or launch the web UI.

---

3. Cloud Authentication

AWS Authentication

gimme-aws-creds should already be installed when you executed pip3 install -r requirements.txt if not, then load up your python virtual environment

pip3 install --upgrade gimme-aws-creds

Configure ~/.okta_aws_login_config and ~/.aws/credentials as described below:

Show AWS Okta config sample

[DEFAULT]
okta_org_url = https://scylladb.okta.com
client_id = 0oab7271m7nRkKmlw5d7
write_aws_creds = True
cred_profile = acc-role
okta_username = <your.name>@scylladb.com
app_url = https://scylladb.okta.com/home/amazon_aws/0oa2uxps59d96E5Cj5d7/272
remember_device = True
preferred_mfa_type = Okta
aws_default_duration = 21600

[cx-cs-lab]
inherits = DEFAULT
aws_rolename = arn:aws:iam::<account>:role/DevOpsAccessRole
cred_profile = DevOpsAccessRole

Add dummy credentials to ~/.aws/credentials:

[DevOpsAccessRole]
aws_access_key_id = dummy
aws_secret_access_key = dummy

Run:

export AWS_PROFILE=DevOpsAccessRole
gimme-aws-creds -p cx-cs-lab

You may wish to add the export to your .bashrc/.zshrc.

---

Directory Structure

.
├── clusters/                   # Auto-generated folders for each cluster
│   └── <cluster-id>/
│       ├── terraform/          # Scylla Cloud Terraform code
│       └── ansible/            # Loader provisioning and config
├── ansible/                    # Shared Ansible playbooks and templates
├── generate_tfvars.py          # Converts variables.yml to tfvars
├── generate_loader_nodes_scripts.py  # Generates SSH and workload scripts
├── scale.py                    # Handles scale in/out using Scylla API
├── get_details.py              # Displays current node info in table
├── get_provider.py             # Determines cloud and version
├── scylla-cloud-operations.sh  # Main lifecycle controller
├── tail_loader_log.sh          # Tails cassandra-stress logs from loader
├── kill_tail.sh                # Kills tailing session from GUI
├── variables.yml               # Unified cluster configuration
└── README.md

Usage

Define Cluster

Edit variables.yml and define one or more clusters under clusters:.

clusters:
  tiny-cluster:
    scylla_api_token: "<your-token>"
    cluster_name: "tiny-cluster"
    ...

Note: Important to change the cluster_name to keep it unique between other users.

Setup

./scylla-cloud-operations.sh setup tiny-cluster

This will:

• Generate tfvars
• Copy templates to cluster specific folders
• Provision infrastructure on Scylla Cloud
• Provision a new VPC with unique CIDR
• Generate Route Tables
• Peer VPCs
• Provision Loader Instances on this VPC

Loader Setup

./scylla-cloud-operations.sh loader_setup tiny-cluster

Stress Testing

./scylla-cloud-operations.sh stresstest tiny-cluster

Kill Load

./scylla-cloud-operations.sh kilload tiny-cluster

Scale

./scylla-cloud-operations.sh scaleout tiny-cluster
./scylla-cloud-operations.sh scalein tiny-cluster

Destroy

./scylla-cloud-operations.sh destroy tiny-cluster

Web UI

A real-time web interface is available via app.py (Flask-based). It lets you:

• Manage clusters
• View logs in real time
• Tail cassandra-stress output from Loaders directly on the browser
• Edit variables.yml from within the UI

Launch the app with:

Activate the Python Virtual Environment and Launch the Backend Server for the Framework using the python -u app.py.

❯ python -u app.py
 * Serving Flask app 'app'
 * Debug mode: on
WARNING: This is a development server. Do not use it in a production deployment. Use a production WSGI server instead.
 * Running on http://127.0.0.1:5000
Press CTRL+C to quit
 * Restarting with stat
 * Debugger is active!
 * Debugger PIN: 625-749-179

Ignore the WARNING, Once you see * Running on http://127.0.0.1:5000 messange, the backend is running successfully. Open up a browser window and go directly to http://127.0.0.1:5000 and have fun :)

Notes

• Use the GUI buttons to test connectivity, provision the cluster infrastructure, run stress test or initial data load, scale out and scale in, kill stress jobs and destroy the cluster.
• CIDRs for Scylla Cloud and Loader VPCs are dynamically generated and isolated.
• Logs and stress tests are streamed via WebSocket.
• Tail logs are streamed from a random loader via SSH.