Install HAProxy (Hardened)

Prerequisites

• An account on the target cloud with permission to launch VM instances (no special IAM roles are required by this product unless stated below — IAM required: none).
• An SSH key pair registered in the target cloud (all access is key-only; password SSH is disabled).
• Instance size ≥ 4 vCPU (pricing floor; 8 vCPU recommended).
• No internet access or external dependency is required at deploy time — the image is self-contained.

AWS

1. Subscribe and launch; open 80/443 in the security group for your frontends.
   Expected result haproxy -v prints the version; stats reachable via SSH tunnel to 8404.

Google Cloud

1. Deploy from the listing; the package opens 80/443 to the network tag (LB posture) and keeps SSH IAP-only.
   Expected result gcloud compute start-iap-tunnel <vm> 8404 → http://localhost:8404/stats renders.
2. Declare a frontend/backend pair in haproxy.cfg, validate (`haproxy -c -f`), then `systemctl reload haproxy`.
   Expected result Traffic to the VM's 80 reaches your backend; stats page shows the backend UP.

Validate

curl -fsS http://127.0.0.1:8404/stats  (on the instance) — the stats page

First boot: Service enabled with the stats frontend on 127.0.0.1:8404; add your frontends/backends to haproxy.cfg.

Secure it

• Restrict SSH (22) to your own IP range in the cloud firewall/security group.
• Open application ports only per the ports table — closed-by-default is deliberate.
• Volume encryption: use your cloud's native volume encryption (enabled by default on most accounts); the image adds no proprietary encryption layer.

Costs & quotas

Software is billed by the marketplace at the listed rate; infrastructure (VM, storage, egress) is billed by your cloud at its standard rates. The recommended size fits default service quotas in most accounts — if you scale out, review your cloud's quota console before launch.

Next: configuration · troubleshooting · security notes