We should introduce the concept of a results directory for a run. Either the user supplies one on the command-line, or one is created automatically (e.g., using subject, date, and time). This directory can later be used to diagnose problems with the tests. And tests then have a place for storing bigger log files etc.
The kaas tests already use a subject-specific directory for placing ephemeral kubeconfig.yaml during provisioning of the test cluster. Note that this directory is not specific to the run (even though it's currently abused also for results). However, we could copy its contents into the run-specific directory before the test is run: (a) so the test doesn't need to look at two different directories, and (b) to preserve its contents as part of the run's documentation.
We should introduce the concept of a results directory for a run. Either the user supplies one on the command-line, or one is created automatically (e.g., using subject, date, and time). This directory can later be used to diagnose problems with the tests. And tests then have a place for storing bigger log files etc.
The kaas tests already use a subject-specific directory for placing ephemeral
kubeconfig.yamlduring provisioning of the test cluster. Note that this directory is not specific to the run (even though it's currently abused also for results). However, we could copy its contents into the run-specific directory before the test is run: (a) so the test doesn't need to look at two different directories, and (b) to preserve its contents as part of the run's documentation.