Manual approval before continuing a job

[clement-uspace]

What's the use case?

In some cases I'd like to add a "manual approval" step to an op. The rationale behind it is that at the end of a run, the results would be published to external entities, and we want to manually check for correctness before that happens.

Ideas of implementation

I'd like to be able to mark an op as "needing approval". There are two ways to interpret this, I'm not sure which is clearer :

• the approval must be done before the op run. In my example case, I would mark the "publish" op as needing approval.
• the approval must be done after the op is run, and before the downstream ops are run. In my example case, I would mark the "compute_results" op as needing approval.

Either way, I'd like to receive an alert when a job is ready for approval, with a "Approve" or "Continue execution" button somewhere.

Additional information

Link to my question in Slack: thread Link to other question in Slack: thread Similar feature in Prefect: https://medium.com/the-prefect-blog/needs-approval-184f2512a3cf

Message from the maintainers

Impacted by this issue? Give it a 👍! We factor engagement into prioritization.

[sryza]

Thanks for the suggestion @clement-uspace . We will keep this in mind in our next prioritization cycle.