Hi, I think what you're looking is:
https://issues.apache.org/jira/browse/NIFI-8385 https://github.com/apache/nifi/pull/4952 This way when a bulletin is generated "because of" a flow file, the flow file UUID would be part of the bulletin payload. It'd mean that in addition to your external ID, storing the flow file UUID would be useful to match things together but I think it'd get you closer to your expectations. HTH, Pierre Le ven. 21 mai 2021 à 11:23, Geir Klingenberg <[email protected]> a écrit : > Hi, > > We are using Nifi as a piece of a larger processing system. Files (and > thus flowfiles) are related to objects that have an ID external to Nifi. To > be able to track objects, flowfiles are given the external ID as a flowfile > attribute. When a processor fails the flowfile is routed to a flow that > updates the monitoring system with the external ID and the error message of > our making. > > But the root cause of the failure, described by the bulletin, is "lost". > That is, we log it using the SiteToSiteBulltinReportingTask, but we have to > manually match the bulletin with our recorded error based on time stamp. > > Do you think it would be possible / good idea to add the bulletin to the > flowfile as part of the Nifi core somehow (AbstractProcessor?), so that any > processor that fail automatically adds a standard attribute to the flowfile > causing the error? > > Thanks, > Geir Klingenberg >
