How HULK Turned Manual Ops into a Productized Cloud Platform

Introduction

The original piece was published in 2013 by the HULK private‑cloud team in CSDN’s Programmer magazine, discussing platform‑centric construction and the enduring relevance of operational tooling and productization.

HULK is now moving toward intelligence, and this article traces its growth from a "zygote" to a mature platform.

First Upgrade: Operations Tooling

Early on, with only two IDC sites and a limited number of servers, operations were largely manual. As projects and server counts exploded, the team adopted automation to handle repetitive tasks.

Following a low‑cost, ready‑to‑use principle, they leveraged open‑source solutions instead of building from scratch. Examples include a custom YUM‑based package manager for rapid installation and scaling, Puppet for configuration management, and SaltStack for batch server control.

Monitoring was split into four layers—system, application, project‑logic, and user‑experience—using tools such as Nagios, Cacti, Ganglia, and Zabbix, supplemented by internally developed utilities. An alarm hierarchy (email warning, SMS alert, aggressive SMS) reduced noise while ensuring critical issues were addressed promptly.

Second Upgrade: Operations Productization

With sufficient tooling in place, the team shifted to a product mindset, naming the platform HULK (the "Green Giant") to symbolize growth on the shoulders of a giant.

The platform introduced a structured request flow, turning ad‑hoc operations into standardized services. Core services (Web, RDB, NoSQL) form the first layer, while generic foundational services constitute the second layer, all designed to reduce communication overhead and accelerate delivery.

A notable feature is the "Project Incubator," which, after a brief input of business name and expected QPS, automatically provisions resources such as MySQL, MongoDB, Redis, web servers, and associated management tools, dramatically shortening the trial‑and‑error phase.

To simplify user interaction, the traditional rigid approval workflow was relaxed; administrators can compile requests directly, avoiding unnecessary re‑submissions.

Conclusion

Operations are critical to a company's success, yet traditional models can stifle technical growth and innovation. By embracing broad‑scope technology, product thinking, and ownership culture, HULK demonstrates how operations can evolve into a platform that not only supports but also accelerates development, delivering superior user experiences.