Tencent SNG Operations: Business Profiling for Capacity Planning, Activity Modeling, and Multi‑Region Deployment

As operations engineers, we often face questions such as how performance changes across device models, how to handle resource waste after holiday traffic spikes, how to identify core modules among hundreds of services, and how to plan multi‑region deployments.

Tencent SNG solves these issues by constructing business portraits that combine module performance data (CPU, memory, I/O), routing relationships, packet captures, and activity metrics to generate four key models.

Capacity Model: Automated load testing across different configurations and versions produces TPS thresholds for each device type, enabling fine‑grained capacity management based on TPS rather than traditional CPU usage.

Activity Model: Historical activity data and downstream module changes are recorded to predict future capacity growth during events, reducing manual effort and optimizing costs.

Core‑Link View: By aggregating routing, packet‑capture, and call‑graph data into a three‑layer (access → logic → storage) call chain and filtering by call frequency and traffic, a concise core service graph is produced for alarm correlation, root‑cause analysis, and architecture optimization.

SET Planning Model: For high‑availability, low‑latency services, business portraits enable multi‑region distribution using a single‑write, multi‑read strategy, visualizing core links, synchronizing data across regions, and defining measurable capacity indicators.

The models are continuously refined: after each activity, real‑world data is compared to predictions, and discrepancies over 20% trigger deeper analysis and model enhancements.

Ultimately, business portraits encapsulate years of operational experience into automated models that drive capacity management, activity forecasting, core‑link visualization, and multi‑region planning, helping teams solve complex operational challenges efficiently.