@nuxeo-nifi, What I would try to implement as a quick solution is: 1. Configure your ValidateRecord (or even maybe try ValidateCSV) so it identifies when your Records from your CSV are not valid. 2. From ValidateRecord, you have 3 possible queues: failure --> which you might want connect to an alert system, like PutEmail for example. valid --> which you might want to connect to your further processing. invalid --> what you are actually looking for :). Here, you can use an InvokeHTTP to call NiFi's REST API and stop your ValidateRecord Processor. In this way, if a single message was rejected, your entire flow will be stopped... this is actually not the best way to do things but if this is your project requirement, this is what you should do.   2a. From ValidateCSV, you have 2 possible queues: valid --> which you might want to connect to your further processing. invalid --> what you are actually looking for :). Here, you can use an InvokeHTTP to call NiFi's REST API and stop your ValidateCSV Processor. In this way, if a single message was rejected, your entire flow will be stopped... this is actually not the best way to do things but if this is your project requirement, this is what you should do. 3. If you are using this Flow in a so to say Streaming mode (you get files every second), you should modify ValidateRecord to run every 5 seconds or every 2 seconds (or something like that) so you have time to stop your processor using InvokeHTTP. If you leave it by default on Run Schedule 0 sec, you will process some additional messages before being able to stop your processor. Documentation: NiFi Rest API: https://nifi.apache.org/docs/nifi-docs/rest-api/index.html NiFi ValidateRecord: https://nifi.apache.org/docs/nifi-docs/components/org.apache.nifi/nifi-standard-nar/1.20.0/org.apache.nifi.processors.standard.ValidateRecord/index.html NiFi InvokeHTTP: https://nifi.apache.org/docs/nifi-docs/components/org.apache.nifi/nifi-standard-nar/1.20.0/org.apache.nifi.processors.standard.InvokeHTTP/index.html NiFi ValidateCSV: https://nifi.apache.org/docs/nifi-docs/components/org.apache.nifi/nifi-standard-nar/1.20.0/org.apache.nifi.processors.standard.ValidateCsv/index.html How To ValidateRecord: https://community.cloudera.com/t5/Community-Articles/Using-Apache-NiFi-to-Validate-that-Records-Adhere-to-a/ta-p/247299     ... View more

@nuxeo-nifi, the processors you are referring to do not belong to any NiFi Version (Cloudera or Open-Source), meaning that they were built in house, specially for you and your project. In this case, you would need to speak to those who have developed those processors and identify the application logic. Once you have that, you can use PutMail to send email notifications and InvokeHTTP to do the other actions.   I assume that from your processors you have has a failed connection queue, which might be linked to an PutMail Processor, in which you define whatever you want to be send as notification   In case of no failures, you can link the success queue out of your nuxeo processor and into InvokeHTTP and perform the call you require. For that, make sure that all your certificates are in place and allow connection between the systems. Otherwise, you won't be able to use InvokeHTTP and you would have to find another solution, like a script. ... View more

@danielhg1285, While the solution provided by @SAMSAL seems to be better for you and more production ready, you could also try the below things. This might work if you are using a stable statement all the time and if are not restricted to see the exact INSERT Statement but rather see the values trying to be inserted. - Shortly after RetryFlowFile, you can add an AttributesToJSON processor and manually define all the columns which you want to insert in the Attributes List Property. Make sure that you use the attribute name from your FlowFile (sql.args.N.value) in your correct order and you set Destination = flowfile-content. In this way, you will generate a JSON File with all the columns and all the values which you have tried to insert but failed. - After AttributesToJSON, you can keep your PutFile to save your file locally on your machine, hence opening it whenever and wherever you want 🙂 PS: This is maybe not the best solution, due to the following reasons, but it will get you started on your track: - You will need to know how many columns you have to insert and each time a new column will be added you will have to modify your AttributesToJSON processor. - You will not get the exact SQL INSERT/UPDATE Statement, but a JSON File containing the column-value pair, which can easily be analyzed by anybody. ... View more