Thanks for the patience and the questions.
Replies below and in the RFC
Why do we need this?
security tools are not always idempotent, hence a workflow that failed for whatever reason cannot always be restarted with the same results
deduplication enrichers save things to a database, if you fail at the consumer stage everything then is a duplicate but has not been reported
the ability to restart a workflow from the last successful stage is a core component of any workflow engine
this is the first step at implementing retryable steps
Is it something urgent?
No, future looking for now, necessary after v1.0
How can we prevent the storage backend from being spammed with huge amounts of data?
I have yet to see a security tool that provides huge numbers of data without the tool itself dying trying to export it. We can design for a scale of low Mb.
How is it enabled thoughout a pipeline?
Every single component takes --in and --out parameters, part of the component definition.
These parameters are currently paths.
In the future they should be urls
What are the new methods that will be added? What are their signatures?
Thanks for the patience and the questions. Replies below and in the RFC
this is the first step at implementing retryable steps
No, future looking for now, necessary after v1.0
I have yet to see a security tool that provides huge numbers of data without the tool itself dying trying to export it. We can design for a scale of low Mb.
Every single component takes
--in
and--out
parameters, part of the component definition. These parameters are currently paths. In the future they should be urlsupdated the RFC.
implementation detail, in the rfc
just the SDK