Errors related to uploading artifacts while tracking runs are silent

[polyaxon-team]

Current behavior

From slack:

Another question about logging metrics to a run through a local jupyter notebook. After our conversation on Nov. 30th. ☝️, things were working fine. However recently the dashboard has stopped displaying metrics again, and I'm seeing weird behaviour in polyaxon. I don't know what has changed.
Looking for advice since I'm out of troubleshooting ideas. Details in the thread...
Here's code I have that recreates the problem:

from polyaxon import tracking

tracking.init(
    owner="owner",
    project="project-name",
    name="test_run",
    run_uuid=None,
    is_new=True
)
tracking.set_run_event_logger()

tracking.log_text(name="some_text_metric", text="some text")

for step in range(1, 100):
    tracking.log_metric(name="some_step_metric", value=step/2, step=step)

tracking.log_succeeded()
tracking.end()

After using some debugging using:

from polyaxon.logger import configure_logger

configure_logger(verbose=True)
...

It turns out that :

Thank you for that command there. After looking at the logging from that, I realized that my polyaxon cli host had switched from the url of the gateway deployed on our cluster to https://cloud.polyaxon.com/.
After some tests, it looks like logging metrics through cloud.polyaxon.com causes the issues I was seeing with artifacts. When I switched the polyaxon host to the url of our gateway, then the dashboard started correctly displaying metrics.

Enhancement

As suggested by the user, the upload is happening in a thread, API errors (404/401/403) should show to help the user debug issues:

Any chance that you'd update the code to provide a useful error message when someone tries this?