Seqera Platform Feature Requests

Currently we have to run GET against individual workflows which can saturate the API. It would be more efficient to run batch requests, e.g. instead of for workflowId in workflows: GET workflow/${workflowId} We could do: GET workflowsBatch?workspaceId=12345678?params=["abcd", "efghi"] This would reduce API burden and not hit rate limits. This could also go for other endpoints, e.g. participants, pipelines etc.

It would be great to have the ability to create access tokens with custom-scoped permissions. Right now, tokens inherit all the permissions of the user who creates them, which can be risky in some situations. Determining specific permissions for each token would make things more secure and flexible.

When retrieving the audit logs data (from the /admin/auditlogs endpoint) there doesn't seem to be a way of filtering the returned data in any way, which means we have to grab all the audit log data and then filter the returned data. Can we add some filtering (especially on dates) so that I can, for example, only return audit log data via the API for the previous day?

Last weekend, we had a rare anomaly in which all Nextflow workflows silently "stalled". I've written code to identify workflows in this condition using the Seqera Platform API . Next I want to cancel and resume the workflows. Cancel is straightforward and it's also easy to Resume via the Seqera Platform/Tower web UI by selecting Resume from the kebab menu on the right-hand side. But what is a good way to do this via the Seqera Platfom API? We use the /workflow/launch endpoint to initiate workflows. When a workflow is cancelled, Seqera Platform/Tower should have all the information to immediately resume the workflow since it was passed by the initial API POST request. As I understand the API, the workflow/launch endpoint has a Boolean resume attribute and I see no alternative. It seems rather long-winded to query the API and use those values to repopulate all the payload attributes for the workflow/launch endpoint, set resume to True and send all this back to the endpoint. Is this the recommended approach or is there a better alternative?