Help Center/ Application Performance Management/ FAQs/ Usage FAQs/ How Does APM Collect Probe Data?
Updated on 2024-03-05 GMT+08:00
View PDF
Share

How Does APM Collect Probe Data?

Collecting Data

Application Performance Management (APM) collects application data through probes. Probes use the bytecode enhancement technology to trace resources and generate call data. The ICAgent obtains and processes the call data. Then, all the data is reported to and displayed on APM. The procedure is as follows:

Collected Data

APM collects service tracing data, resource information, resource attributes, memory monitoring information, and call request KPI data, but does not collect your personal data. The collected data is used only for APM performance analysis and fault diagnosis, and is not used for any commercial purposes. The following table lists the details.

Data Type

Collected Data

Transmission Mode

Storage Mode

Data Purpose

Storage Period

Tracing data

Tracing span data

Transmission through HTTPS encryption and Access Key ID/Secret Access Key (AK/SK) authentication

Project-based isolated storage

Query and display at the tracing frontend

Configurable (7 days at most). The data will be deleted upon expiration.

Call request KPI data

Call initiator address, receiver address, API, duration, and status

Transmission through HTTPS encryption and AK/SK authentication

Project-based isolated storage

Calculation of transaction call KPI metrics, such as throughput, TP99 latency, average latency, and error calls, drawing of application topologies, and display at the frontend

7 days. The data will be deleted upon expiration.

Resource data

Service type, service name, creation time, deletion time, node address, and service release API

Transmission through HTTPS encryption and AK/SK authentication

Project-based isolated storage

Query and display at the resource library frontend

7 days. The data will be deleted upon expiration.

Resource attributes

System type, system startup event, number of CPUs, service executor, service process ID, service pod ID, CPU label, system version, web framework, JVM version, time zone, system name, collector version, and LastMail URL

Transmission through HTTPS encryption and AK/SK authentication

Project-based isolated storage

Query and display at the resource library frontend

7 days. The data will be deleted upon expiration.

Memory monitoring data

Memory usage, used memory, maximum memory, remaining memory, memory threshold-crossing time, and memory monitoring configurations

Transmission through HTTPS encryption and AK/SK authentication

Project-based isolated storage

Query and display at the resource library frontend

7 days. The data will be deleted upon expiration.

APM Resource Overhead

The CPU usage of each probe is less than 5% and used memory is about 250 MB.

Parent topic: Usage FAQs