Simon
Simon
After PR #1847 displaying the SQL error in Python remains as TODO.
I'm in favor of it: any arbitrary key-value mapping should be in its own object.
Yes, this needs back-end support.
Currently not planning to do this, pipeline-uuid is unique identifiable, making it (tenant_id, name) in my opinion just makes it longer. I would say keep open for future potentially.
Sounds good, let's tackle this step-wise :+1: This sounds like a good start
New API spec: - Checkout the PR: https://github.com/feldera/feldera/pull/2071 - Run the pipeline-manager (note that it will delete all existing pipeline-related tables) - Go to: http://localhost:8080/swagger-ui/
Looks good, it has all the components, just needs to be made slightly more general and a couple of fixes to fit into the current database implementation architecture.
I would recommend keeping the database table and its associated functions quite generic, then we can later on decide further what constitutes an event. It might be best to keep...
It needs to be adjusted for post-pipeline-rework Feldera. I'll coordinate the work with @swanandx . My envisioned goal is that we use the same strategy for cluster status history and...
Yes, I agree it only makes sense for cases where there is a bound amount of time it should take (for example that particular state transition). We should make a...