A smart DevOps assistant powered by local AI to manage and interact with your Kubernetes clusters. Ask questions in natural language and get real-time information about your pods, logs, and more!
- Natural Language Interface: Interact with your Kubernetes cluster using plain English. No more complex
kubectlcommands to memorize! - Pod Management: Easily list running pods, check their status, and get essential details.
- Log Inspection: Instantly fetch and view logs for any specific pod.
- AI-Powered: Utilizes a local Ollama model to understand your requests and translate them into the right actions.
- Extensible: Built with Java, making it easy to add new tools and capabilities.
- Privacy-Focused: Your data and cluster information stay on your machine, as it uses a local AI model.
Follow these steps to get the DevOps Co-Pilot up and running on your local machine.
- Java (JDK 11 or higher): Make sure you have a compatible Java version installed.
- Maven: For building the project and managing dependencies.
- Docker: Required for running a local Kubernetes cluster (like Minikube or Kind).
- kubectl: Configured to connect to your Kubernetes cluster.
- Ollama: You must have Ollama installed and running. You can download it from ollama.com.
-
Clone the repository:
git clone [https://github.com/your-username/DevOpsCoPilot.git](https://github.com/your-username/DevOpsCoPilot.git) cd DevOpsCoPilot -
Pull the Ollama model: The agent needs a model to function. We recommend a function-calling model like
llama3orphi3.ollama pull llama3
-
Build the project: Using Maven:
mvn clean package
-
Run the agent:
java -jar target/devops-copilot-0.1-SNAPSHOT.jar
Once the agent is running, you can interact with it from its command-line interface.
- List all pods in the default namespace:
"show me all the pods"
- Get logs for a specific pod:
"can you get me the logs for the pod named 'api-gateway-xyz-123'?"
- Check the status of containers present:
"what's the status of the containers present?"