Application Performance Monitoring¶
Application Performance Monitoring (APM), based on standardized protocols such as OpenTracing, achieves full-stack performance analysis through distributed tracing technology. It directly correlates server metrics, application logs, and user access data to enable:
- End-to-end full path tracing
- Code-level fault point localization
- Correlation analysis of resource consumption and performance bottlenecks
Optimal Deployment¶
Deploy the DataKit agent on each application server to achieve:
- Host resource metrics collection (CPU, memory, disk, etc.)
- Application and operating system log capture
- Distributed call chain tracing
After processing the data locally with DataKit, all data is uniformly reported to Guance for correlation computation.
Use Cases¶
Core Features¶
-
View service list, performance metrics, and invocation relationships between services
-
Query and analyze all collected trace data using tools like flame graphs
-
View historical trends and distribution of similar errors within traces
-
View number of online services, P90 service response time, maximum service impact time, and other metrics
-
View application runtime metrics and real-time display of call relationships and efficiency
-
Application Performance Detection
Detect abnormal trace data by configuring application performance monitors
Storage Billing¶
Application Performance Monitoring billing counts the number of trace_id within the current workspace, using a tiered pricing model.
For details on data storage policies, please refer to Data Storage Policy.
For more billing rules, please refer to Billing Methods.