Deploy Basics

Goal: Deploy AgenticGoKit applications to production environments with proper configuration and monitoring.

Prerequisites

Complete tools-basics.md to understand MCP tool integration.

What is Deployment?

Deployment involves:

• Building production-ready binaries
• Configuring environment variables
• Containerizing applications
• Managing secrets securely
• Setting up monitoring

1) Validate configuration for production

agentcli validate

Expected: All configuration validates without errors before deployment.

2) Build production binary

# Build optimized binary
go build -ldflags="-s -w" -o agentapp .

This creates a production binary with debug information removed for smaller size.

3) Create production Dockerfile

FROM golang:1.21-alpine AS build
WORKDIR /app
COPY go.mod go.sum ./
RUN go mod download
COPY . .
RUN go build -ldflags="-s -w" -o agentapp .

FROM alpine:latest
RUN apk --no-cache add ca-certificates
WORKDIR /app
COPY --from=build /app/agentapp .
COPY --from=build /app/agentflow.toml .
EXPOSE 8080
CMD ["./agentapp"]

4) Build Docker image

docker build -t agenticgokit-app:v1.0.0 .

Expected: Image builds successfully with optimized layers.

5) Configure environment variables

Create production environment setup:

# Required API keys
$env:OPENAI_API_KEY = "your-production-api-key"

# Optional: Custom configuration file
$env:AGENTFLOW_CONFIG = "agentflow.prod.toml"

# Optional: Database connection for memory
$env:DATABASE_URL = "postgres://user:pass@db:5432/agents"

6) Run containerized application

docker run --rm `
  -e OPENAI_API_KEY=$env:OPENAI_API_KEY `
  -e DATABASE_URL=$env:DATABASE_URL `
  -p 8080:8080 `
  agenticgokit-app:v1.0.0

Expected: Application starts and responds on port 8080.

7) Production configuration example

Create agentflow.prod.toml:

[agent_llm]
provider = "openai"
model = "gpt-4o"
max_tokens = 4000
temperature = 0.7

[agent_memory]
provider = "pgvector"
connection = "${DATABASE_URL}"

[agent_logging]
level = "info"
format = "json"
output = "stdout"

[agent_tracing]
enabled = true

8) Health monitoring

Add health check endpoint to your application:

// Add to main.go
http.HandleFunc("/health", func(w http.ResponseWriter, r *http.Request) {
    w.WriteHeader(http.StatusOK)
    w.Write([]byte("OK"))
})

Test health endpoint:

curl http://localhost:8080/health

9) Common deployment patterns

Cloud platforms

• AWS: Use ECS, EKS, or Lambda
• Google Cloud: Use Cloud Run, GKE
• Azure: Use Container Instances, AKS

Container orchestration

• Docker Compose: Multi-service local deployment
• Kubernetes: Production-grade orchestration
• Docker Swarm: Simplified container clustering

Security considerations

• Store API keys in secure secret management
• Use TLS for all external communications
• Implement proper access controls
• Regular security updates

Next Steps

• Explore advanced deployment patterns in Examples.md
• Learn about production monitoring in ErrorHandling.md
• Review security best practices in Providers.md

Verification checklist

• [ ] agentcli validate succeeded
• [ ] Production binary built successfully
• [ ] Docker image built without errors
• [ ] Environment variables configured
• [ ] Container runs and responds to requests
• [ ] Health check endpoint accessible
• [ ] Production configuration file created