Currently, all agents except the Python agent support the log_level option.

We'd like to add this to central config to make it easier to increase the log level on-the-fly to troubleshoot issues without having to restart the application.

Open Questions

• Does exposing a log_level option make sense for Python? Currently, the log level is configured via the framework's log config. Would adding an Elastic APM specific log_level option interfere with that? How do other agents, such as the .NET agent handle this that can also be configured via the framework's logging config?
• Is it feasible to make this option dynamic for all agents?
• Should we align on values?
  • Would it be strange for .NET or Python user, for example, to use different log_levels in central config as opposed to the ones used in the framework's logging config?
  • Do we need to enable users to set the log_level across apps monitored by different agents?

Status

Spec: #332 ✅