Weekend Load balanced Application and DB Cluster Deployment with HAProxy

I have seven freshly racked Ubuntu bare-metal servers on the same private network and I want them production-ready before the weekend is over.

Architecture to build
• 3 servers become the application tier: you will install the exact PHP and Node.js versions I will provide, configure them behind Cloudflare’s load balancer (active-active) and stand up a three-node Redis cluster.
• 4 servers form the data tier: MySQL and Postgres must run in a highly-available cluster on these four machines. All application traffic—internal and external—should arrive through HAProxy, which you will configure for read/write awareness and fail-over.

Monitoring & dashboards
Grafana (with Prometheus, node exporter, MySQL, Postgres and Redis exporters) needs to be in place on every box, charting CPU, memory, uptime, LAN/WAN traffic per interface, plus live “top queries” panels for MySQL, Postgres and Redis.

What I will hand you
– The exact version list for PHP, Node.js, MySQL and Postgres.
– SSH access to each server and Cloudflare credentials.

Success means that by Monday Morning at 10:30 in India I can:
1. Push a PHP or Node app and reach it through the Cloudflare VIP.
2. Write to either database engine and watch HAProxy route correctly.
3. See all requested metrics on the Grafana dashboards.

If you have recent hands-on experience with Ubuntu server automation, HAProxy tuning, database clustering and Grafana observability, this should be a fast, clean weekend build.