How Intelligent Operations and Observability Transform Cloud‑Native Environments

Background

On August 19‑20, 2022, the GOPS Global Operations Conference was held in Shenzhen, where Guance Cloud’s chief evangelist Wu Yakun presented the topic “Intelligent Operations and Observability in the Cloud Era.”

Traditional Operations Challenges

Traditional operations relied on simple three‑tier architectures and manual “three‑hammer” fixes—restart, reinstall, replace hardware. As workloads moved to virtualization, cloud, and micro‑services, these methods became insufficient because service call chains grew longer and root‑cause identification grew harder.

Intelligent Operations in the Cloud Era

In the cloud‑native era, massive data volumes require new strategies to monitor every layer and quickly locate issues. The primary goal of monitoring is to ensure business stability by enabling rapid fault detection and recovery.

Unified Data Collection and Analysis

Effective analysis of massive data demands unified storage, correlation, and analysis. Isolated monitoring systems lead to fragmented data and high overall fault rates. By consolidating data, teams can trace a problem through its entire call chain, identify the responsible component, and resolve issues efficiently.

Observability Directions

Observability encompasses monitoring, AIOps, and deeper analysis. It involves collecting low‑density metrics (CPU, memory) alongside high‑density signals (traces, logs) and using them to drive actions. Unified data enables SLO‑driven dashboards that automatically flag violations and support KPI assessment.

Open‑Source Tools: DataKit and DataFlux Func

Guance Cloud open‑sourced the DataKit collector, which gathers data from diverse sources and forwards it to a backend platform. DataFlux Func provides a programmable open‑source sandbox for integrating cloud billing, sensor data, and custom scripts, facilitating extensive data enrichment.

SLO‑Based Monitoring

By packaging metrics into Service Level Objectives (SLOs) such as 99.9% availability, the platform can automatically deduct budgets on violations, offering clear visibility into system health and supporting performance‑based assessments.

Conclusion

The talk emphasizes that unified data collection, correlation, and analysis are essential for intelligent operations and observability, enabling faster fault isolation, cost optimization, and better alignment between development and operations.