Currently, one have to use some non-public (and probable not intended) API, from the pipeline script:

String getCurrentFlowNodeID() {
    // Types explanation:
    //   CpsThread.current() : CpsThread
    //   .head               : FlowHead (final field access, no accessor method present)
    //   .get()              : FlowNode
    //   .getId()            : String
    return CpsThread.current().head.get().getId()
}

Background:

We need a business-level view on certain customized pipeline aspects, but with the ability to dig lower into technical foundation.
Therefor we need the ability to map between these customized aspects and (at least) the FlowNode where they occurred.
We are "afraid" that the current approach using non-public API is not intended and may be changed or even get removed.

FYI: we iterate post-mortem through builds/[num]/workflow/*.xml files to instantiate a simple data model view on the pipeline, which is then enriched with data from another bunch of our own *.xml files, where the customized business aspects are stored. Both kinds of XML data contain FlowNode IDs, so that they can be mapped together.