Updated on 2022-07-13 GMT+08:00

Overview

O&M Challenges

In the cloud era, there are more and more applications under the distributed microservice architecture. The number of users also increases explosively, facing various application exceptions. In traditional O&M mode, metrics of multiple O&M systems cannot be associated for analysis. O&M personnel need to check application exceptions one by one based on experience, resulting in low efficiency, costly maintenance, and poor stability.

When there are massive quantities of services, O&M personnel face two major challenges.

  • Large distributed applications have complex relationships, making it difficult to analyze and locate problems. Specifically, O&M personnel face problems such as how to ensure normal application running, and quickly locate faults and performance bottlenecks.
  • Users choose to leave due to poor experience. O&M personnel fail to detect and track services with poor experience in real time, and cannot diagnose application exceptions in a timely manner, severely affecting user experience.

Introduction to APM

Application Performance Management (APM) monitors and manages the performance of cloud applications in real time. APM provides performance analysis of distributed applications, helping O&M personnel quickly locate and resolve faults and performance bottlenecks.

As a cloud application diagnosis service, APM has powerful analysis tools. It displays the application status, call processes, and user operations based on topologies, tracing, and transaction analysis, so that you can quickly locate and resolve faults and performance bottlenecks.

Figure 1 APM architecture
  1. Access APM: You can access APM by creating an Identity and Access Management (IAM) agency and implementing Access Key ID/Secret Access Key (AK/SK) authentication.
  2. Data collection: APM can collect data about applications, basic resources, and user experience from Java probes, and Istio mesh in non-intrusive mode.
  3. Service implementation: APM supports topologies, tracing, and transaction analysis.
  4. Service expansion:
    • Application Operations Management (AOM) monitors application O&M metrics in real time while APM quickly diagnoses application performance bottlenecks through topologies and tracing.
    • Cloud Performance Test Service (CPTS) implements association analysis and generates performance reports after APM identifies performance bottlenecks.
    • Based on the historical metric data learned using intelligent algorithms, APM associates metrics for analysis from multiple dimensions, extracts the context data of both normal and abnormal services for comparison, and locates root causes through cluster analysis.

Advantages

Connects to applications without having to modify code, and collects data in a non-intrusive way. Data comes from:

  • Java probe: Collects service call data, service inventory data, and call KPI data in non-intrusive mode based on the pinpoint open-source project.
  • Istio mesh: Collects service call data, resource information, and call KPI data in non-intrusive mode through the Kubernetes platform.

Delivers high throughput (hundreds of millions of API calls), ensuring premium experience.

Analyzes root causes using AI-powered threshold detection and machine learning based on historical baseline data.

Opens O&M data query APIs and collection standards, and supports independent development.