Specifying RTN Codes to Ignore

# CADS Return Code Ignore Feature - Request Summary

## 📋 Request Overview

We request the addition of a feature to **ignore specific return codes** in stream job steps and general job steps in IBM SPSS Collaboration and Deployment Services (CADS).

---

## 🔴 Current Issues

### Issue 1: Lack of Flexibility

- When a return code is not 0 in a job step, the job step is automatically treated as "Failed"

- The overall job status also becomes "Failed"

### Issue 2: Unable to Distinguish from Errors

- Cannot distinguish between intentional non-zero return codes (e.g., for data-driven branching) and actual errors

- Results in false positives in operations

---

## 🎯 Why This Feature is Needed

### Enabling Data-Driven Processing Flows

**Business Requirements:**

- In data analysis, it is necessary to select different processing paths based on data content and state

- Example: Send email or create model (100) when data count exceeds threshold, do nothing (0) when below threshold

**Current Constraints:**

- When return code 100 is returned, the job is treated as failed

- This is not actually an error, just a data condition being met

- **Result**: Normal business logic is misidentified as "Failed"

**Specific Impact:**

- Cannot distinguish between intentional control codes (100, 200, etc.) and unexpected errors (1, 2, etc.)

- Operations staff must check logs every time to determine if it's a true error

- Increased workload due to unnecessary alert responses

- Risk of overlooking true errors

---

## 💡 Usage Scenario Example

### Scenario: Processing Branch Based on Data Volume

Processing branches can be implemented using Script API.

```python

# Modeler Script

record_count = data.valueAt(0, 0)

if record_count == 0:

    modeler.script.exit(100)  # No data (warning level)

elif record_count > 1000:

    modeler.script.exit(200)  # Threshold exceeded (information level)

else:

    modeler.script.exit(0)    # Normal processing

```

**Current Behavior:**

- When return codes 100 or 200 are returned, the job step is treated as failed

- Even though it's not an error, the entire job has a failed status

**Expected Behavior:**

- Set return codes 100 and 200 in the ignore list

- Job is treated as successful in all cases

- Select appropriate processing path in the next step using conditions like `completion_code == 100`

---

## ✅ Request Details

### Feature Request

**Addition of a feature to configure "Return Codes to Ignore"**

### Specific Behavior

- When return code is **100** or **200**:

  - Do **not** set the job step to Failed

  - Overall job status does not become Failed

  - Proceed normally to the next step

### Expected Benefits

1. **Flexible Processing Flow Control**: Enable branching based on data content

2. **Error Clarification**: Distinguish between actual errors (unexpected return codes) and intentional control codes

3. **Improved Operational Efficiency**: Reduce unnecessary investigation work due to false positives

---

## 🎯 Implementation Image

### Configuration Items (Expected)

- Add "Return Codes to Ignore" field in job step configuration screen

- Allow multiple values separated by commas: `100, 200, 300`

### Operation Logic

```

IF Return Code == 0 THEN

    → Success

ELSE IF Return Code IN Ignore List THEN

    → Success (Ignored Code)

ELSE

    → Failed

END IF

```

---