Feature Overview (aka. Goal Summary)  

Improve the installer logging to specifically identify error causes as much as possible to better understand what’s going on in the installation process.

Goals (aka. expected user outcomes)

Users should be able to understand the root cause of the installer errors as much as possible.

Requirements (aka. Acceptance Criteria):

The errors detected provide enough information in the log message to narrow the cause down as much as possible rather than just plain generic messages that take a lot of time to debug.

Option to dump ALL logging messages to the serial console to be scraped by the cloud platform (Azure, AWS, GCP)

Background

Errors from the Installer have been hard to identify and understand and needs to be refined, possibly with error codes for each error.

In the cloud after an install has failed, the cloud resources are cleaned up. This makes it impossible to login to the failing machines afterwards.

However all log messages sent to the serial console can be scraped by the cloud platforms. With an option to make a noisy install the Managed Openshift SRE teams will be able to get the data to quantify the installation success SLA. 

Documentation Considerations

Initially there are no documentations requirements for this feature.