Python Runtime Metrics (OpenTelemetry)

Overview

The Python OpenTelemetry integration allows you to collect runtime metrics from your Python applications instrumented with OpenTelemetry.

Setup

See the dedicated documentation for Python Runtime Metrics via OpenTelemetry.

The following metrics are emitted as cumulative counters by the OpenTelemetry Python SDK. To convert them to per-second rates in Datadog, apply the Delta To Rate processor in your OpenTelemetry Collector pipeline:

  • process.cpu.time
  • process.context_switches
  • process.runtime.cpython.gc_count

Data Collected

Metrics

process.cpu.time
(gauge)		Number of seconds of CPU time used by the process
Shown as second
process.cpu.utilization
(gauge)		CPU utilization of the process as a fraction of available CPU per core
Shown as percent
process.context_switches
(gauge)		Number of context switches for the process
Shown as invocation
process.runtime.cpython.gc_count
(gauge)		Number of objects in garbage collection generations
Shown as resource
process.thread.count
(gauge)		Number of threads in the process
Shown as thread
process.memory.usage
(gauge)		Amount of physical memory in use by the process
Shown as byte

Support

Need help? Contact Datadog support.