dapr-agents/quickstarts/05-multi-agent-workflow-k8s

Run Multi agent workflows in Kubernetes

This quickstart demonstrates how to create and orchestrate event-driven workflows with multiple autonomous agents using Dapr Agents running on Kubernetes.

Prerequisites

• Python 3.10 (recommended)
• Pip package manager
• OpenAI API key
• Kind
• Docker
• Helm

Configuration

1. Create a .env file for your API keys:

OPENAI_API_KEY=your_api_key_here

Install through script

The script will:

1. Install Kind with a local registry
2. Install Bitnami Redis
3. Install Dapr
4. Build the images for 05-multi-agent-workflow-dapr-workflows
5. Push the images to local in-cluster registry
6. Install the components for the agents
7. Create the kubernetes secret form .env file
8. Deploy the manifests for the agents
9. Port forward the workload-llm pod on port 8004
10. Trigger the workflow for getting to Morder by k8s_http_client.py

Install through manifests

First create a secret from your .env file:

kubectl create secret generic openai-secrets --from-env-file=.env --namespace default --dry-run=client -o yaml | kubectl apply -f -

Then build the images locally with docker-compose:

docker-compose -f docker-compose.yaml build --no-cache

Then deploy the manifests:

kubectl apply -f manifests/

Port forward the workload-llm pod:

kubectl port-forward -n default svc/workflow-llm 8004:80 &>/dev/null &

Trigger the client:

python3 services/client/k8s_http_client.py