이 페이지는 아직 한국어로 제공되지 않으며 번역 작업 중입니다. 번역에 관한 질문이나 의견이 있으시면 언제든지 저희에게 연락해 주십시오.

The Agent is CPU-bound and its CPU usage is correlated with the number of spans received per second.

The Agent buffers unprocessed payloads in memory, so throttling the Agent process because of an insufficient CPU limit can lead to an out-of-memory issue.

Detect out-of-CPU

To monitor CPU usage and detect oncoming out-of-CPU issues, compare the maximum CPU percentage configured for the Agent to the datadog.trace_agent.cpu_percent metric. The datadog.trace_agent.cpu_percent metric is CPU usage in terms of percentage of a core. For example, a value of 50 is half a core, or 200 is two cores.

See the full list of Agent APM metrics.

Resource requirements

A good indicator to calculate adequate resource limits for the Agent is the number of spans received per second, reported in the datadog.trace_agent.receiver.spans_received metric. Based on that metric’s value, follow the table below to choose adequate CPU and memory limits:

Spans per secondCPU (core)Memory (MB)
20000.0535
11 0000.240
32 0000.660
58 000170
130 0002130

Notes:

  • The values are based on Agent 7.39.0 benchmarks.
  • The benchmarks were performed on an AWS c5.2xlarge instance (8 VCPU/ 16GiB RAM)