Installation & Deployment Guide

This guide covers the installation requirements and the different operational modes for RadOps.

Prerequisites

Before running RadOps, ensure the following infrastructure is available:

• Python 3.11+
• Redis: Required for short-term conversation memory (falls back to in-memory if unavailable).
• HashiCorp Vault: Required for secret management (secrets can be set via env vars or config if unavailable).
• Vector Database: One of Weaviate, Chroma, Qdrant, Milvus or Pinecone (for RAG).

Quick Start: Infrastructure via Docker

To quickly spin up the required dependencies (Redis, Vault) and an optional Vector DB (e.g., Weaviate), you can use the following docker-compose.yml.

1. Create docker-compose.yml:

services: # 1. Redis (Required for Memory) redis: image: redis/redis-stack-server:latest container_name: radops-redis ports: - "6379:6379"

# 2. Vault (Required for Secrets)
vault:
  image: hashicorp/vault:latest
  container_name: radops-vault
  ports:
    - "8200:8200"
  environment:
    VAULT_DEV_ROOT_TOKEN_ID: 'my-dev-token'
  cap_add:
    - IPC_LOCK

# 3. Vector DB (Optional - Example: Weaviate)
weaviate:
  image: semitechnologies/weaviate:1.27.0
  container_name: radops-weaviate
  ports:
    - "8080:8080"
    - "50051:50051"
  environment:
    QUERY_DEFAULTS_LIMIT: 25
    AUTHENTICATION_ANONYMOUS_ACCESS_ENABLED: 'true'
    PERSISTENCE_DATA_PATH: '/var/lib/weaviate'
    DEFAULT_VECTORIZER_MODULE: 'none'
    ENABLE_MODULES: ''
    CLUSTER_HOSTNAME: 'node1'
```

2. Run the stack: bash docker-compose up -d

Installation

1. Clone the repository:

   git clone https://github.com/mehrdadrad/radops.git
   cd radops
   

2. Install dependencies: It is recommended to use a virtual environment.

   python -m venv venv
   source venv/bin/activate
   pip install .
   

3. Configuration: Copy the example configuration and adjust it to your environment.

   cp config.example.yaml config.yaml
   cp tools.example.yaml tools.yaml
   cp integrations.example.yaml integrations.yaml
   cp rbac.example.yaml rbac.yaml
   cp server.example.yaml server.yaml
   

   Refer to the config_guide.md, tools_guide.md, integrations_guide.md, and rbac_guide.md for detailed configuration options.

Running Modes

RadOps operates in two distinct modes depending on your use case.

1. Server Mode (Production)

Server Mode is the intended way to run RadOps in a production environment. In this mode, the application:

• Starts the REST API server to handle incoming requests.
• Initializes background schedulers for syncing RAG data sources (e.g., GitHub, Google Drive).
• Logs output to the file specified in config.yaml (e.g., logs/app.log).
• Loads server configuration from config/server.yaml.

Command:

python server.py

Slack Integration:

The Slack integration is managed by src/integrations/slack/slack.py. This script runs a Slack bot that connects to the RadOps server via WebSockets.

For detailed configuration instructions, required scopes, and token generation, please refer to the Slack Integration Guide.

Running the Bot: Ensure server.py is running, then start the Slack bot:

python src/integrations/slack/slack.py

2. Console Mode (Debugging & Testing)

Console Mode provides an interactive command-line interface (REPL). It is strictly designed for development, debugging, and testing.

Use Console Mode to:

• Test new Tool definitions or MCP connections immediately.
• Debug LLM prompts and responses in real-time.
• Verify connectivity to infrastructure (Vault, Redis, Vector DB) without starting the full web server.
• Inspect memory state.

Command:

python console.py