Hi, i'm experiencing a new issue with logging scalars and plots. it seems that something has changed on your side (clearml) and scalar and plots are having trouble getting logged (usin the pro plan of clearml). the result is a full training run without scalars logged (the same code worked perfectly a few days ago). it seems like the issue is not persistent (for example, this morning the logging has succeeded, and after a few hours new tasks fails to log). looking at the attached console log, you can see that the logs order doesn't makes sense (image fetching is at the middle of the log after reporting epoch results...).
another example is the attached image, where logs are ok until epoch 227, but stops reporting afterwards (has completed 400 epochs).
these issues resulting tasks that are stuck under "running" mode although they finished the training, probably due to logs that needs to be done and are not completed (we use "task.flush(wait_for_uploads=True)" at the end of our scripts)
any help regarding these issues would be much appreciated!