Open alstolten opened 1 year ago
Pinging @elastic/elastic-agent (Team:Elastic-Agent)
This likely stopped working in the 8.6 release. What is the use case for turning these off?
These metrics can be the only way for us diagnose performance issues, so having them always be present in the logs helps root cause analysis significantly.
There is still a path in the code that leads to the flag here being disabled:
It looks like we set the YAML configuration to "-"
in the struct tags which means it isn't actually configurable:
This was originally added in https://github.com/elastic/elastic-agent/commit/393b2f018d3e347a750b3dceab55867967ecf338, so it looks like we accidentally disabled the path in the code that allows turning this off based on configuration.
Hey @cmacknz the user wants to disable those because there are many messages from many agents reaching their cluster. They reported to have used the options in the documentation and were wondering why this does not work.
Version: 8.7.1
Operating System: Linux/Windows
Description: Using the
agent.logging.metrics.enabled
oragent.logging.metrics.period
for a stand-alone Elastic-Agent, as described in the reference yaml does not work. The options will not be forwarded to the underlying beats components. Ie they will be started without the necessary-E logging.metrics.enabled=false
command line argument.Working example from
7.17.11
:Steps to Reproduce:
elastic-agent-8.7.1-linux-x86_64
agent.logging.metrics.enabled: false
to the elastic-agent.yml and install/run the agent