Cloud‑Native Infrastructure Practices at Manbang Group’s HuoCheBang

Innovation at internet scale prioritizes speed, mobile‑first user reach, and guarantees of scalability and availability, prompting a shift toward cloud‑native software architecture.

01 Monitoring – Alerts, metrics, distributed tracing, and data visualization form the heart of distributed operations; the unified monitoring platform Galileo ingests over 70 billion data entries daily, storing more than 3 TB of data.

02 Container Cloud – Built on Kubernetes, the container cloud platform Planck abstracts the runtime environment, offering resource allocation and isolation, container deployment and scheduling, certificate and secret management, log archiving, usage tracking, and multi‑data‑center support. It hosts all HuoCheBang services across three centers and eight partitions, running over 4,000 containers.

03 Services – The microservice approach decomposes applications into independently deployable, scalable components, isolating failures and supporting independent evolution. The microservice environment Newton defines contracts for service metadata, governance, resource planning, access control, configuration injection, ACLs, deployment modes, dependency isolation, and fault handling. Integrated with Planck, Galileo, the versioned configuration system Kip, the programming framework Maxwell, and the API gateway Lorentz, Newton provides service registration, discovery, routing, load balancing, fault tolerance, and access control.

More than 40 % of HuoCheBang’s services have migrated to microservices, accounting for roughly 75 % of deployment frequency.

04 DevOps and R&D Management – The Solvay platform enhances efficiency and supports rapid, frequent delivery and operations. Acting as the primary interface with development teams, Solvay integrates other infrastructure systems, R&D management tools, and testing facilities, unifying the delivery workflow around a “Feature” lifecycle and exposing metrics such as change lead time, deployment frequency, change‑failure rate, work‑in‑process, and test‑failure rate.

The infrastructure team is dedicated to providing flexible, self‑service software infrastructure that enables continuous value delivery for development teams.

Author: Wang Ruoyu, Head of Infrastructure, HuoCheBang Technology Department, Manbang Group.