Hi community, when I configure a ClearML client to S3 storage backend instead of the default file server, I assume that the connection to the S3 backend happens directly to the S3 endpoint and not routed through any service from the ClearML services like API server etc. My understanding is that the ClearML client utilizes Boto3 API calls underneath the hood that directly connect to the S3 endpoint using the credential chain locally.

• Does ClearML store any dataset outside of this S3 storage (and local storage) for preview or compression purposes?
• Are any data packets related to model or data versioning routed through the API server? (We are pretty confident it shouldn't, but just want to be sure from a security standpoint)
  We are trying to make the case for the security of the tool with the relevant stakeholders, so would appreciate if there was any sort of documentation link that cited this and we might have missed out. Sincere apologies if we did miss it, and thanks in advance for sharing the answer. 🙂