Skip to main content

Dynamically scale an RDS Instance

important

This guide requires an Upbound control plane instance running UXP v2.0 or later. Upbound SaaS coming soon.

Upbound Crossplane is capable of running Intelligent Controllers, which define AI-augmented functions to perform tasks. This guide walks through a use case for using AI to intelligently scale an AWS RDS database instance.

Prerequisites​

Before you begin make sure you have:

  • An Upbound Account
  • The up CLI installed
  • An Anthropic API key
  • An AWS account

Set up your environment​

Clone the repository upbound/configuration-aws-database-ai to your machine:

git clone git@github.com:upbound/configuration-aws-database-ai.git

This repository contains a control plane project that defines a fully managed AWS database instances. This database instance is implemented with an Intelligent Composition function that scales the database in relation to performance metrics fetched from AWS CloudWatch.

Launch the local UXP cluster​

In the root of the project directory, launch the control plane locally:

up project run --local

Configure credentials and runtime settings​

Create a secret on the control plane which contains an Anthropic API key:

kubectl create secret generic claude \
  --from-literal=ANTHROPIC_API_KEY=<your-api-key> \
  -n crossplane-system

Create a second secret on the control plane which contains credentials for your AWS account:

kubectl create secret \
  generic aws-creds \
  -n crossplane-system \
  --from-file=aws-creds=./aws-credentials.txt

Create a database​

Deploy a network and database:

# Network dependency
kubectl apply -f examples/network-rds-metrics.yaml

# Database with scaling enabled
kubectl apply -f examples/mariadb-xr-rds-metrics.yaml

The database gets created, the control plane periodically fetches performance metrics for it from AWS CloudWatch, and dynamically scales the database size accordingly.

To validate the intelligent scaling system, you can stress test the RDS instance to trigger high CPU utilization and observe AI-driven scaling decisions. The command below performs a stress test to mimic real usage:

# Trigger high CPU load with multiple concurrent MD5 hash computations
for i in {1..8}; do
    mysql \
      --host=your-rds-endpoint.region.rds.amazonaws.com \
      --user=masteruser \
      --password=your-password \
      --default-auth=mysql_native_password \
      --execute="SELECT BENCHMARK(1000000000, MD5('trigger_scaling_$i'));" &
done

Clean up​

Clean up the local control plane to prevent it from continuing to invoke your LLM. Run the following command:

up project stop

Next steps​

Read the concept documentation for Intelligent Controllers to learn more about using AI-powered functions in your function pipelines.