Standardizing Operations and Building a Public‑Cloud CMDB to Accelerate Deployment

Author Li Bin, an operations development engineer, shares the challenges faced when rapid business growth outpaces operational capabilities, such as lack of standards, manual operations, incomplete monitoring, and scattered logs.

To address these issues, the team introduced automation and standardization, establishing consistent naming conventions for applications and log formats, which enable automatic log collection and integration with monitoring, CI/CD, and other systems.

They then built a public‑cloud Configuration Management Database (CMDB) to capture standardized metadata about applications rather than individual hosts, focusing on application management and reducing asset‑centric complexity.

The CMDB supports three lifecycle stages: creation, runtime (adding or scaling hosts), and decommissioning. During creation, a unified entry generates corresponding records in monitoring, release, and service‑registry systems. In the runtime stage, entering an IP adds the host to the CMDB and triggers initialization and code deployment via Ansible playbooks. Decommissioning removes the host from all linked systems.

To further improve efficiency, a self‑service code release platform was upgraded to use OpenResty with Consul KV for dynamic upstreams, exposing APIs for CMDB host addition and node lifecycle management, and allowing customizable release steps via Ansible playbooks with fine‑grained authorization.

In practice, the new workflow reduces a multi‑machine scaling operation during the 618 promotion from roughly 18 minutes to about five minutes, allowing operators to simply add hosts in the CMDB and let the system handle initialization, deployment, and load‑balancer updates automatically.

The overall takeaway is that by standardizing operations, building a cloud‑native CMDB, and automating release processes, operational efficiency dramatically improves, freeing engineers to focus on higher‑value tasks such as continuous integration, reliability, and cost optimization.