Fred

Getting Started

Start the backend

You need :

  • make
  • Docker if you want to run the application in Docker
  • Python 3.12.1 for local development

To try out Fred, you simply need to start the python backend:

cd backend
make run

The equivalent local command is:

python fred/main.py --server.configurationPath ./config/configuration.yaml --server.baseUrlPath /fred

Fred backend expects a valid OPENAI token in the OPENAI_API_TOKEN environment variable, and a kubeconfig. Your kubeconfig can point to any kubernetes cluster including k3d, kind, minikube. Using OpenAi is not mandatory, you can use a local ollama server or AzureOpenAi as well.

A great starting point is to try it out with kooker, or simply with a local minikube.

From there, Fred will start collecting and generating data and save it under ~/.fred/fred-backend-cache (as set in the resource/configuration-dev.yaml file).

If you do not want at all to interact with a remote or development K8 cluster nor perform some openai requests, get in touch with contributors to get a ready-to-use zip archive that you can simply unzip in ~/.fred/. You will then work in offline mode.

You can use the LOG_LEVEL environment variable.

LOG_LEVEL=DEBUG python fred/main.py --server.configurationPath ./config/configuration.yaml --server.baseUrlPath /fred

The Swagger UI is accessible at http://<host>:<port>/docs. Follow the host and port of your configuration file. You will have the documentation of the APIs and buttons to test it (Try it out).

You will see something like:

Configuration File

Fred backend accepts a configuration file where you can fine-tune many of its internal services. Here is a self documented version:

# In development mode it is handy to work with data files instead of connecting to remote databases.
database:
  type: csv
  csv_files:
    # Can be absolute paths or relative paths to the main
    energy_mix: './services/cluster_consumption/data/simulated_energy_mix.csv'
    carbon_footprint: './services/cluster_consumption/data/simulated_cluster_consumption_gco2_sep_to_feb.csv'
    energy_footprint: './services/cluster_consumption/data/simulated_cluster_consumption_wh_sep_to_feb.csv'
    financial_footprint: './services/cluster_consumption/data/simulated_cluster_consumption_usd_sep_to_feb.csv'

kubernetes:
  kube_config: '~/.kube/config'
  aws_config: '~/.aws/config' # Optional, needed for aws EKS clusters.
  # Timeout settings for the client
  timeout:
    connect: 5  # Time to wait for a connection in seconds
    read: 15    # Time to wait for a response in seconds

ai:
  # Timeout settings for the client
  timeout:
    connect: 5  # Time to wait for a connection in seconds
    read: 15    # Time to wait for a response in seconds
  default_model:
    #provider: "ollama"
    #name: "llama2"
    provider: "openai"
    name: "gpt-4o"
    #provider: "azure"
    #name: "fred-gpt-4o"
    api_version: "2024-05-01-preview"
    temperature: 0.0
  leader:
    name: "Fred"
    class_path: "leader.leader.Leader"
    enabled: true
    model: {}
  services:
    - name: "kubernetes"
      enabled: true
      model: {}
  agents:
    - name: K8SOperatorExpert
      class_path: "agents.kubernetes_monitoring.k8s_operator_expert.K8SOperatorExpert"
      enabled: true
      mcp_servers:
        - name: k8s-mcp-server
          transport: sse
          url: http://localhost:8081/sse
          sse_read_timeout: 600 # 10 minutes. It is 5 minutes by default but it is too short.
      model: {}
    - name: "GeneralistExpert"
      class_path: "agents.generalist.generalist_expert.GeneralistExpert"
      enabled: true
      model: {}
    - name: "DocumentsExpert"
      class_path: "agents.documents.documents_expert.DocumentsExpert"
      enabled: true
      categories:
        - "eco-conception"
      settings:
        chunk_size: 512
        chunk_overlap: 64
        knowledge_flow_url: "http://localhost:8111/knowledge/v1" 
      model: {}
      
# Where to save fred produced resources like Essentials or Score
# and external resources like Kubernetes Workload descriptions
dao:
  type: "file"  # Currently the only one supported
  base_path: "~/.fred/dao-cache"
  max_cached_delay_seconds: 300  # Cache delay in seconds. Use 0 for no cache or a negative value for limitless cache.

# Where to store user feedback
feedback:
   type: postgres
#  db_host: fred-postgres
#  db_port: 5432
#  db_name: fred_db
#  user: admin
#  password: Azerty123_

# Enable or disable the security layer
security:
  enabled: false
  keycloak_url: "http://fred-keycloak:8080/realms/fred"

# KEYCLOAK
keycloak:
  server_url: "https://localhost:8080"
  realm_name: "fred"
  client_id: "fred"

# CONTEXT STORAGE CONFIGURATION
context_storage:
  type: "local"
  options:
    path: "~/.fred/context-store"

Online Parameter

usage: main.py [-h] [--server.address SERVER_ADDRESS] [--server.port SERVER_PORT] [--server.baseUrlPath SERVER_BASE_URL_PATH] [--server.configurationPath SERVER_CONFIGURATION_PATH]
               [--server.logLevel SERVER_LOG_LEVEL]

Fred microservice

options:
  -h, --help            show this help message and exit
  --server.address SERVER_ADDRESS
                        Specify the address of the server
  --server.port SERVER_PORT
                        Specify the port of the server
  --server.baseUrlPath SERVER_BASE_URL_PATH
                        Specify the base url for the API entrypoints
  --server.configurationPath SERVER_CONFIGURATION_PATH
                        Specify the path of the configuration used
  --server.logLevel SERVER_LOG_LEVEL
                        Specify the log level of the server

Start the UI

Requirements

make, curl, node, nvm, and docker are required to build and run the project on your laptop.

For docker read [installation guide for Ubuntu](https://docs.docker.com/engine/install/ubuntu/, https://docs.docker.com/engine/install/linux-postinstall/)

make run

Which is equivalent to:

npm run dev

Go to the indicated localhost url and use the UI with the credentials fred/fred.

Prev
Backend Deployment Guide
Next
How to use Fred