You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Add a --with-diagnostics flag to generate a diagnostics bundle if there is any error when installing the Elastic Agent.
Describe a specific use case for the enhancement or feature:
If any error happens during the install and enroll process the Elastic Agent will clean up any leftover, if the installation happened it'll remove the installation folder together with all the logs. Logs which are valuable when trying to diagnose issues during install/enroll.
What is the definition of done?
When an installation using the --with-diagnostics flag fails, it should produce a diagnostics bundle containing, if available:
the installation logs. It's all the output shown to the user when invoking elastic-agent install ...
the Elastic Agent and its components logs
the installed Elastic Agent state
if possible the current configuration for each component
Ideally, the process controlling the installation would request a diagnostics from the agent daemon, if it isn't possible, it'd collect the necessary information
The text was updated successfully, but these errors were encountered:
Ideally, the process controlling the installation would request a diagnostics from the agent daemon, if it isn't possible
If installation fails, there likely is no running Elastic Agent daemon. If we've gotten that far, installation has actually succeeded.
We could write the install logs to disk in the directory the install command was run from. This saves a user from have to do this via copy paste. When installations are done via some orchestration tool like Ansible I'm not sure if this helps, because it adds another thing for users to have to remember to cleanup and pull off the machine.
We could write the install logs to disk in the directory the install command was run from. This saves a user from have to do this via copy paste. When installations are done via some orchestration tool like Ansible I'm not sure if this helps, because it adds another thing for users to have to remember to cleanup and pull off the machine.
We could write to the tmp directory of the system or, as you suggested on slack, on the folder where the agent binary is, so when the folder is deleted after the installation, those logs are also deleted.
Describe the enhancement:
Add a
--with-diagnostics
flag to generate a diagnostics bundle if there is any error when installing the Elastic Agent.Describe a specific use case for the enhancement or feature:
If any error happens during the install and enroll process the Elastic Agent will clean up any leftover, if the installation happened it'll remove the installation folder together with all the logs. Logs which are valuable when trying to diagnose issues during install/enroll.
What is the definition of done?
When an installation using the
--with-diagnostics
flag fails, it should produce a diagnostics bundle containing, if available:elastic-agent install ...
Ideally, the process controlling the installation would request a diagnostics from the agent daemon, if it isn't possible, it'd collect the necessary information
The text was updated successfully, but these errors were encountered: