More control over resuming suspended nodes

[markterm]

Summary

Enable the resume command/API to resume just nodes with arguments matching a parameter, and have the ability to set a suspended node to failure.

Motivation

In the case of a workflow triggering asynchronous execution, eg a third party service that uploads data to S3. At the end of the execution we would want to unsuspend just that specific node, and if the execution fails then we would want to be able to set it to failure. Currently this would generally be done via a pod that is polling the third-party service, but that's not optimal from a performance perspective.

Proposal

Add extra parameters to unsuspend API with an optional argument matching expression and an optional error message that would set the Suspend node to failure and record the message against the node.

[simster7]

Sounds like a good idea. Are you interested in submitting a PR for this?

[markterm]

Yes, I'll be creating it within the next couple of weeks.

[markterm]

I've created #1904