This topic describes the performance monitoring feature of the Full-stack Observability application.

Benefits

Service development complicates the deployment architecture of applications. More factors are involved, including the microservice architecture, multi-cloud environment, cross-region deployment, and multi-architecture machines. Continuous performance data collection introduces the temporal dimension to performance data collection. The following table describes the comparison between continuous performance data collection and single-host performance data collection.

Scenario	Single-host performance data collection	Continuous performance data collection
Compare the performance of a service between the test environment and production environment	Not supported	Supported
Compare the performance of a service between version 1.0 and version 2.0	Not supported	Supported
Compare the performance of a service between an ARM server and an AMD server	Not supported	Supported
Compare the performance of a service between the China (Beijing) region and the China (Hangzhou) region	Not supported	Supported

Collection mode

Simple Log Service Performance Monitoring is an open and comprehensive platform that supports continuous performance data analysis. Developers can collect data to Performance Monitoring by using different programming languages.

Logtail supports the pull and push data collection modes.

In pull mode, Logtail can automatically pull performance data from programs that are written in specific programming languages.
- Collect performance data from Go programs
- Collect performance data from Go programs in a Kubernetes cluster.
In push mode, Logtail can receive performance data that is reported by specific platforms over transmission protocols such as Pyroscope.