Skip to main content

​
Deployment Overview

Solace Agent Mesh offers flexible deployment options designed to meet different operational requirements, from local development to production-grade infrastructure. Understanding these options helps you choose the right approach for your specific environment and scale needs.

​
Deployment Strategies

Agent Mesh supports multiple deployment strategies, each optimized for different use cases:

​
Development Deployment

During development, simplicity and rapid iteration are key priorities. The Agent Mesh CLI provides a streamlined way to run your entire project as a single application, making it easy to test changes and debug issues locally. Run your entire project: 
sam run
This command starts all configured components together, providing immediate feedback and allowing you to see how different agents interact within your mesh. Run specific components: 
# Single agent or workflow
sam run <agent or workflow config file path>

# Multiple components together
sam run <agent config file path> <workflow config file path>
The development setup automatically loads environment variables from your configuration file (typically a .env file at the project root), eliminating the need for complex environment management.

​
Production Deployment

Production deployments require different considerations than development environments. You need:
  • Reproducible builds - Consistent runtime environments across deployments
  • Scalable infrastructure - Ability to handle varying loads
  • Robust monitoring - Comprehensive observability and health checks
  • High availability - Fault tolerance and automatic recovery
  • Security - Proper secrets management and encrypted communication
We recommend containerization for production deployments:
  • Docker - For single-node deployments or simpler orchestration needs
  • Kubernetes - For multi-node, scalable, highly available deployments

​
Deployment Architecture Patterns

​
Monolithic Deployment

Deploy all Agent Mesh components as a single container or process. This approach:
  • Simplifies initial deployment and management
  • Reduces operational complexity
  • Works well for smaller deployments or testing
  • Limits independent scaling of components
Example Docker command: 
FROM solace/solace-agent-mesh:latest
WORKDIR /app
COPY . /app
CMD ["run", "--system-env"]

​
Microservices Deployment

Deploy components as separate containers that communicate through the Solace event broker. This approach:
  • Enables independent scaling of components
  • Provides better fault isolation
  • Allows granular resource management
  • Supports rolling updates per component
  • Increases operational complexity
Key considerations:
  • Reuse the same Docker image across components
  • Customize startup commands per deployment
  • Ensure shared storage configuration is identical
  • Configure health checks for each component

​
Hybrid Deployment

Combine monolithic and microservices patterns by grouping related agents together while separating high-traffic or resource-intensive components:
  • Core platform and orchestrator in one deployment
  • Specialized agents in separate deployments
  • Gateways as independent services

​
Infrastructure Requirements

​
Compute Resources

Minimum requirements vary based on deployment size: Development:
  • 2 CPU cores
  • 4 GB RAM
  • 10 GB storage
Production (per node):
  • 4+ CPU cores (ARM64 or x86_64)
  • 16+ GB RAM
  • 50+ GB SSD storage
  • Additional capacity per concurrent agent (~625 MiB RAM, 175m CPU)

​
Persistence Layer

Database (Required):
  • PostgreSQL 17+ for session storage
  • Managed services recommended for production (AWS RDS, Azure Database, Cloud SQL)
  • SQLite acceptable for development only
Object Storage (Required):
  • S3-compatible API for artifact storage
  • Supported services: AWS S3, Azure Blob Storage, Google Cloud Storage, MinIO
  • Filesystem storage acceptable for development only

​
Network Connectivity

Outbound (Required):
  • Solace event broker (Cloud or self-hosted)
  • LLM provider endpoints (OpenAI, Anthropic, Azure OpenAI, etc.)
  • Container registry access
  • Identity provider (IdP) for authentication
Inbound (For web access):
  • HTTPS/443 for web UI and API
  • TLS termination recommended at ingress layer

​
Deployment Options Comparison

FeatureDockerKubernetes
Setup ComplexityLowMedium to High
ScalingManualAutomatic (HPA)
High AvailabilityRequires external orchestrationBuilt-in
Rolling UpdatesManual processNative support
Health MonitoringExternal tools neededNative probes
Secrets ManagementEnvironment variablesKubernetes Secrets
Best ForSmall deployments, single nodeProduction, multi-node, enterprise

​
Core Components

Understand what gets deployed in a typical Agent Mesh installation: Platform Services:
  • Agent Mesh Core - Central orchestration and configuration management (175m CPU / 625 MiB RAM)
  • Deployer - Dynamic agent and gateway deployment (100m CPU / 100 MiB RAM)
  • Web UI Gateway - User interface and API endpoints (included in core)
Agents:
  • Main Orchestrator - Routes requests and coordinates agent interactions
  • Specialized Agents - Custom or pre-built agents for specific tasks
  • Each agent: ~175m CPU / 625 MiB RAM
Optional Components:
  • Additional Gateways - REST API, WebSocket, or custom integrations
  • Workflows - Declarative multi-step automation
  • Custom Agents - Your own agent implementations

​
Environment Configuration

All deployment methods use environment variables for configuration. Key variables include: 
# Required - Solace Event Broker
SOLACE_BROKER_URL="wss://your-broker.messaging.solace.cloud:443"
SOLACE_BROKER_USERNAME="your-username"
SOLACE_BROKER_PASSWORD="your-password"
SOLACE_BROKER_VPN="your-vpn"

# Required - LLM Configuration
LLM_SERVICE_ENDPOINT="https://api.openai.com/v1"
LLM_SERVICE_API_KEY="sk-..."
LLM_SERVICE_PLANNING_MODEL_NAME="openai/gpt-4"
LLM_SERVICE_GENERAL_MODEL_NAME="openai/gpt-4"

# Required - Security
SESSION_SECRET_KEY="your-random-secret-key"

# Production - Database
DATABASE_URL="postgresql://user:password@host:5432/sam"

# Production - Object Storage
ARTIFACT_STORAGE_TYPE="s3"
ARTIFACT_STORAGE_S3_BUCKET="your-bucket-name"
ARTIFACT_STORAGE_S3_REGION="us-east-1"
AWS_ACCESS_KEY_ID="your-access-key"
AWS_SECRET_ACCESS_KEY="your-secret-key"

# Production - Queue Configuration
USE_TEMPORARY_QUEUES="false"

# Optional - Web UI Configuration
CONFIG_PORTAL_HOST="0.0.0.0"
FASTAPI_HOST="0.0.0.0"
FASTAPI_PORT="8000"

​
Next Steps

Choose your deployment method:

Docker Deployment

Deploy Agent Mesh using Docker containers for single-node or simple orchestration scenarios

Kubernetes Deployment

Deploy Agent Mesh on Kubernetes for production-grade, scalable infrastructure

Production Best Practices

Learn security, monitoring, and operational best practices for production deployments

Build docs developers (and LLMs) love

Get started for freeTalk to us