[Feature Request]: Be able to configure VFS configurations for a Web Service / Asynchronous Web Service

[DBiLeSi]

What would you like to happen?

Consider the following scenario: I have a HOP Server running in a Docker container. On this server, I want to create multiple reports for different individuals and upload them via VFS to these users. I can create a workflow and an asynchronous web service for each report and trigger them accordingly.

However, currently, when I want to upload Report A to GSC Bucket A using VFS Configuration A and Report B to GSC Bucket B with a different VFS Configuration B, I am unable to do so. This limitation arises because changing the configuration requires a server restart.

I noticed in the roadmap that you plan to implement VFS configuration per project. However, this would not resolve the issue, as there is only one project per server, if I am not mistaken.

My suggestion is to either allow VFS configuration for each asynchronous web service or enable multiple projects per server with a mechanism for specifying which project an asynchronous web service should use. Alternatively, you could allow VFS configuration for environments, along with a way to designate the environment for an asynchronous web service

Issue Priority

Priority: 3

Issue Component

Component: Hop Server, Component: VFS, Component: Other

[hansva]

In the current release, we have the option to add multiple destinations for Azure blob storage.
We plan to integrate these for each VFS type in due time.

The only thing they do not allow is to dynamically change them, they get refreshed/loaded on project load/switch.
This has been done because creating these virtual file systems is very resource intensive and creating them at runtime (per pipeline/workflow) would add significant overhead.

[DBiLeSi]

Thank you for the quick reply.

I am mostly interested in the GCS. Just to clarify, if you implement the multiple GCS configuration like you have for Azure Blob Storage, and are in a situation where I have one server and want to upload two different reports to two different GCS buckets, could I save the first export in something like gcs_a://bucket_a and the second in gcs_b://bucket_b without restarting the server?

[hansva]

I will leave this open to track that use case. When you add them to the folder, we might have to add an endpoint to refresh the server metadata. If the configuration is included in the request this should work.