Hi @<1523703961872240640:profile|CrookedWalrus33> , metrics are considered as scalers, logs, plots, and the experiment objects themselves that are saved in the backend databases.

You must be reporting some very metric heavy experiments 🙂

@<1523701070390366208:profile|CostlyOstrich36>
Well, I doubt that it is the case in my situation.
Is there any API where I can get read the metrics per experiment?

I did found this:

from clearml.backend_api.session.client import APIClient
client = APIClient()
client.events.get_task_metrics(tasks=['some_task_id'], event_type='log')

but it returns an empty 'metrics' object

Thanks @<1523701070390366208:profile|CostlyOstrich36> .
What are considered as experiment objects?

the experiments themselves 🙂

Imagine if you have very large diffs or very large (several mb) configuration files logged into the system - this is sitting somewhere in the backend on a database