Deployment Overview

Flexo can be deployed using two main approaches, each suited for different needs and scenarios.

Deployment Approaches

Approach 1: Build-time Configuration

With this approach, you bake your agent configuration into the container image during build time. Environment variables are still injected at runtime for security.

graph LR
    config1[agent.yaml + tools] --> build[Build Image]
    build --> reg1[Push to Registry]
    reg1 --> deploy1[Deploy Container]
    env1[.env variables] -.-> deploy1

    classDef config stroke:#0288d1
    classDef env stroke:#388e3c
    classDef action stroke:#ffa000

    class config1 config
    class env1 env
    class build,reg1,deploy1 action

Best for:

Stable configurations
Version-controlled agent behavior
Immutable deployments
Faster container startup

Approach 2: Runtime Configuration

With this approach, you use a base image and inject both configuration and environment variables at runtime. This allows for more flexible configuration management.

graph LR
    base[Base Image] --> reg2[Push to Registry]
    reg2 --> deploy2[Deploy Container]
    config2[agent.yaml + tools] -.-> deploy2
    env2[.env variables] -.-> deploy2

    classDef config stroke:#6C8EBF
    classDef env stroke:#63825A
    classDef action stroke:#DD7E6B

    class config2 config
    class env2 env
    class base,reg2,deploy2 action

Best for:

Dynamic configuration updates
Environment-specific settings
Testing different configurations
Separate configuration management

Deployment Process

The general deployment process follows these steps: