Loading

Salesforce Nonprofit Cloud: Data Processing Engine run fails with RECIPE_EXECUTION_ERROR_JOB_KILLED_DUE_TO_INACTIVITY

Fecha de publicación: Jul 20, 2026
Descripción

This article covers the failure where a Data Processing Engine (DPE) definition in Nonprofit Cloud terminates instead of completing, and dependent records such as Transaction Journals are never created.

The exact error is:

This run of the Data Processing Engine definition couldn't process the data. DataflowInstance <DataflowInstance ID> Cancelled. Job killed due to inactivity. Error Code: SYSTEM_ERROR, Error Type: RECIPE_EXECUTION_ERROR_JOB_KILLED_DUE_TO_INACTIVITY. The run was terminated. Try running the definition again.

The failure is most commonly reported for the Accounting Subledger (ASL) job and Run Subledger processing. It is triggered when the DPE processes a data volume that exceeds the memory allotted to the underlying CRM Analytics recipe engine. It has been observed in Enterprise, Performance, Unlimited, and Developer Edition orgs with Nonprofit Cloud, and occurs in both sandbox and production (including full-copy sandboxes). The user running the definition must hold the permission set that grants Data Pipelines / DPE execution.

Customers describe this as:

  • "batch job returns failure without a reason"
  • "ASL job failing since <date>"
  • "subledger job doesn't create Transaction Journal records — job killed due to inactivity"
Solución

Work through the causes in order. Most "job killed due to inactivity" failures are Cause 1.

Cause 1: The CRM Analytics recipe memory factor for the org is too low for the data volume.

  1. Confirm the failure signature. Go to Setup > Data Processing Engine > (your definition) > Run History, or Setup > Monitor Workflow Services > Batch Jobs, and open the failed DataflowInstance. Verify the Error Type reads exactly RECIPE_EXECUTION_ERROR_JOB_KILLED_DUE_TO_INACTIVITY.
  2. Reduce the volume the recipe must hold in memory in a single run. Open the DPE definition (DataProcessingEngineDefinition). On each source node, add or tighten Filter criteria so the run processes a bounded set — for example, filter by a date range, a single Account, or a status value — instead of all records at once.
  3. Re-run the definition (Setup > Data Processing Engine > (your definition) > Run Now). A run that completes on a smaller filtered set confirms the failure is memory/volume related, not a data or permission defect.
  4. To process the full volume, request a CRM Analytics recipe memory-factor increase for the affected org. This is a per-org limit adjustment performed by Salesforce Support. Open a case and provide:
    1. The exact org ID (production or sandbox)
    2. The DPE definition name
    3. One failed DataflowInstance ID
    After the increase is applied, re-run the definition without the temporary filters and confirm it completes.

Cause 2: The running user lacks Data Processing Engine / Data Pipelines execution access.

  1. Go to Setup > Users > Permission Sets and confirm the user who launches the job is assigned a permission set granting Data Pipelines and Data Processing Engine access.
  2. Confirm Data Pipelines is enabled (Setup > Data Processing Engine > Get Started / Enable). Note: an access problem typically surfaces as USER_ERROR_ACCESS_MISSING rather than the inactivity error. If you see that instead, resolve access first, then re-run.

Cause 3: A required source object has all fields removed from Data Sync.

  1. Go to Setup > Analytics > Data Manager > Data Sync and Connections and verify each object used by the definition has at least one synced field. Do not remove all fields from an object; if the object is not needed, disconnect the object instead.
  2. Re-run the definition after correcting the sync configuration.

Confirm resolution: Re-run the Accounting Subledger / Run Subledger definition (Setup > Data Processing Engine > (your definition) > Run Now) and verify the run completes with status Success and that the expected Transaction Journal records are created for the processed accounts.

Número del artículo de conocimiento

005389388

 
Cargando
Salesforce Help | Article