Skip to main content

AWS Marketplace Deployment

This guide covers the Marketplace-based In-VPC deployment flow for AxonFlow Enterprise. The current Marketplace package is built around an AWS CloudFormation template that launches the core enterprise services inside your AWS account.

At a high level, the deployment includes:

  • AxonFlow Agent
  • AxonFlow Orchestrator
  • customer portal and customer portal UI when enabled
  • PostgreSQL on RDS
  • ALB, ECS Fargate services, and supporting AWS infrastructure
  • optional Prometheus and Grafana components depending on deployment options

What This Deployment Path Is For​

Use the AWS Marketplace path when:

  • your security or procurement process prefers Marketplace procurement
  • you want AxonFlow to run entirely in your AWS account
  • you need the customer portal, enterprise governance, and controlled deployment operations

This path is especially relevant for regulated or latency-sensitive deployments where keeping traffic inside your own VPC matters.

Prerequisites​

Before you launch the Marketplace stack, have these ready:

  • an AWS account with permissions for CloudFormation, ECS, RDS, EC2 networking, IAM, and ALB resources
  • one VPC with at least 2 public and 2 private subnets across different availability zones
  • the AxonFlow enterprise license provided during commercial onboarding
  • ACM certificate details if you want HTTPS termination through the ALB from day one

What the CloudFormation Template Produces​

The current Marketplace package uses the ECS Fargate template under the AWS Marketplace deployment assets. The template exports outputs including:

  • AgentEndpoint
  • OrchestratorServiceName
  • CustomerPortalEndpoint when the portal is deployed
  • CustomerPortalServiceName when the portal is deployed
  • CustomerPortalUIServiceName when the UI is deployed

Those outputs are the first things platform teams should collect after a successful stack launch.

  1. Subscribe to the AxonFlow Marketplace offering.
  2. Continue to configuration and launch the CloudFormation stack.
  3. Choose the VPC, subnet, and sizing parameters that match your environment.
  4. Wait for the stack to reach CREATE_COMPLETE.
  5. Record the outputs, especially the Agent and portal endpoints.
  6. Verify health, ECS service counts, and basic portal access before loading real traffic.

Minimum Verification Checklist​

Confirm stack status​

aws cloudformation describe-stacks \
  --stack-name axonflow-production \
  --region YOUR_AWS_REGION \
  --query "Stacks[0].StackStatus"

Inspect stack outputs​

aws cloudformation describe-stacks \
  --stack-name axonflow-production \
  --region YOUR_AWS_REGION \
  --query "Stacks[0].Outputs" \
  --output table

Check Agent health​

AGENT_ENDPOINT=$(aws cloudformation describe-stacks \
  --stack-name axonflow-production \
  --region YOUR_AWS_REGION \
  --query "Stacks[0].Outputs[?OutputKey=='AgentEndpoint'].OutputValue" \
  --output text)

curl -sf "${AGENT_ENDPOINT}/health"

Check ECS service health​

aws ecs describe-services \
  --cluster axonflow-production-cluster \
  --services axonflow-agent-service axonflow-orchestrator-service \
  --region YOUR_AWS_REGION \
  --query 'services[].{service:serviceName,running:runningCount,desired:desiredCount}' \
  --output table

If the customer portal was deployed in this stack, validate its endpoint as well before onboarding platform admins.

Commercial And AWS Cost Notes​

The AWS Marketplace path covers the AxonFlow commercial subscription plus the AWS infrastructure you run in your own account. Confirm current subscription terms in AWS Marketplace or with AxonFlow before procurement, and estimate AWS costs for RDS, NAT gateways, ALB, ECS Fargate, logs, and optional monitoring components based on your selected region and availability-zone layout.

What to Validate Before Production Traffic​

  • the Agent health endpoint responds
  • ECS running counts match desired counts
  • the portal endpoint is reachable if deployed
  • database and secret wiring succeeded
  • the enterprise license is valid in the deployed environment
  • security groups allow the expected Agent to Orchestrator and service-to-database traffic

This is where many enterprise rollouts either gain confidence or accumulate hidden debt. Treat this as a real production readiness gate, not a "quick smoke test."

Common Deployment Mistakes​

  • treating the Marketplace deployment as a black box and not recording outputs
  • assuming the portal is always deployed, even when its conditional resources are disabled
  • validating only ALB reachability instead of Agent health and ECS service health
  • skipping the day-zero portal login and license verification step
Last updated on