Use Execution Status to Skip Downstream Calls After Upstream Failures

When a validation or preprocessing node fails or produces no results, downstream nodes should not execute (and waste API calls). Use n8n expressions to check the previous node's execution status and short-circuit the workflow.

Real-world example: A workflow validates incoming data, enriches it via an API, then sends it to an LLM. If validation fails, the enrichment and LLM calls should not run.

// Code node: "Check Upstream Status"
// Place this before any expensive API call

const prevStatus = $('Validate Data').isExecuted;
const validationResult = $('Validate Data').first().json;

if (!prevStatus || validationResult.valid === false) {
  // Return empty to stop this branch, or return an error object
  return [{
    json: {
      skipped: true,
      reason: 'Upstream validation failed',
      errors: validationResult.errors || ['Unknown validation error'],
      apiCallsSaved: 2  // enrichment + LLM
    }
  }];
}

// Validation passed -- continue with the data
return [{
  json: {
    skipped: false,
    data: validationResult.data
  }
}];
```text
A simpler approach using the IF node with execution expressions:

```yaml
IF node configuration:
  Condition: Expression
  Value 1: {{ $('Validate Data').first().json.valid }}
  Operation: equals
  Value 2: true

  True branch:  --> [Enrich via API] --> [Send to LLM]
  False branch: --> [Log Failure] --> [End]
```text
You can also use `$prevNode` for simpler chains:

```javascript
// Expression in an IF node
// Check if the previous node produced any output items
{{ $prevNode.first() !== undefined && $prevNode.first().json.error === undefined }}
```text
```yaml
Cost impact example:
  Workflow: Validate --> Enrich ($0.01) --> LLM ($0.03)
  Invalid input rate: 30%
  Daily executions: 1000

  Without status check: 1000 x ($0.01 + $0.03) = $40/day
  With status check:    700 x ($0.01 + $0.03) + 300 x $0.00 = $28/day
  Savings: $12/day = $360/month
```text
This pattern ensures you only spend money on API calls that will actually produce useful results.

**Related:** [Use Structured Output (JSON Mode) for Parseable Responses](../ai-and-llm-integration/01-use-structured-output-json-mode-for-parseable-responses.md) | [Configure Payload Size and Binary Data Mode for Large Files](../performance-and-large-files/01-configure-payload-size-and-binary-data-mode-for-large-files.md)