Skip to content

API_Testing_Migration.md

Latest commit

 

History

History
245 lines (188 loc) · 6.18 KB

API_Testing_Migration.md

File metadata and controls

245 lines (188 loc) · 6.18 KB

API_Testing.py Migration Notes

Changes Made

The API_Testing.py script has been updated to work with the new Temporal REST API instead of the AWS API Gateway endpoint.

What Changed

1. Endpoint URL

Before (AWS):

url = "https://<random_id>.execute-api.us-east-2.amazonaws.com/v1/execution"

After (Temporal):

url = "http://localhost:8000/v1/execution"

2. Response Processing

Before (AWS Response Format):

{
  "executionArn": "arn:aws:states:...",
  "startDate": "2026-02-24T...",
  ...
}

After (Temporal Response Format):

{
  "workflow_id": "person-number-uuid",
  "run_id": "temporal-run-id",
  "status": "started"
}

3. Results Dictionary Structure

Before:

results_dict[name] = {
    "base": num1,
    "add": num2,
    "sum": num_sum,
    "uuid": exe_num,
    "result": response.json()
}

After (Enhanced with extracted fields):

results_dict[name] = {
    "base": num1, 
    "add": num2, 
    "sum": num_sum, 
    "uuid": exe_num, 
    "result": response_data,
    "workflow_id": response_data.get("workflow_id", "N/A"),
    "run_id": response_data.get("run_id", "N/A"),
    "status": response_data.get("status", "unknown")
}

What Stayed The Same

✅ Request Payload Format - Still uses the same nested JSON structure with escaped strings:

data = '{' + f'\\"personName\\": \\"{name}\\", \\"baseNumber\\": \\"{num1}\\", \\"additionalNumber\\": \\"{num2}\\"' + '}'
body = '{' + f'"input": "{data}", "name": "{exe_num}", "stateMachineArn":"..."' + '}'

✅ Test Data - Still reads from names.txt

✅ Random Number Generation - Same ranges (0-99 for base, 200-299 for additional)

✅ Output File - Still writes to results.json

✅ Validation Logic - Same expected sum calculation

Testing Results

Test Execution

python API_Testing.py

Results

  • ✅ Total workflows started: 124/124 (100%)
  • ✅ Errors: 0
  • ✅ Duplicate handling: 6 duplicate names handled by UPSERT (same as before)
  • ✅ Execution time: ~0.5 seconds to start all workflows (vs ~1 hour sequential on AWS)

Sample Output

{
    "Emma": {
        "base": 33,
        "add": 272,
        "sum": 305,
        "uuid": "5583d7b2-11a7-11f1-bad3-4ef5e7f1951a",
        "result": {
            "workflow_id": "person-number-25de6018-2393-4fbc-8ba2-bb2b4be6e7ed",
            "run_id": "019c90b9-30d4-7ce0-8667-03f9ce2a4221",
            "status": "started"
        },
        "workflow_id": "person-number-25de6018-2393-4fbc-8ba2-bb2b4be6e7ed",
        "run_id": "019c90b9-30d4-7ce0-8667-03f9ce2a4221",
        "status": "started"
    }
}

Verification

1. Check Workflows Started

python3 -c "import json; r = json.load(open('results.json')); print(f'Total: {len(r)}, Success: {sum(1 for v in r.values() if \"workflow_id\" in v)}')"

Output: Total: 124, Success: 124

2. Check Database After Completion (wait ~35 seconds)

psql -d temporal_migration_test -c "SELECT COUNT(*) FROM persons;"

3. Verify Specific Results

psql -d temporal_migration_test -c "SELECT person_name, assigned_number FROM persons WHERE person_name IN ('Emma', 'Liam', 'Olivia') ORDER BY person_name;"

4. Check Temporal UI

Visit http://localhost:8233 to see all workflow executions in the Temporal Web UI.

Functional Equivalence

Aspect AWS Implementation Temporal Implementation Match?
Input Format Nested JSON with escaped strings Same ✅
Test Data names.txt Same ✅
Number Generation Random 0-99, 200-299 Same ✅
Calculation base + additional Same ✅
Execution API Gateway → Step Functions REST API → Temporal ✅
Database DynamoDB PostgreSQL ✅
Results Format JSON with execution details JSON with workflow details ✅
Duplicate Handling DynamoDB UPSERT PostgreSQL UPSERT ✅
Workflow Timing 30 second wait Same ✅

Backward Compatibility

To switch back to AWS (not recommended):

# Uncomment the AWS URL
url = "https://<random_id>.execute-api.us-east-2.amazonaws.com/v1/execution"

# And remove the Temporal-specific field extraction
results_dict[name] = {
    "base": num1,
    "add": num2,
    "sum": num_sum,
    "uuid": exe_num,
    "result": response.json()
}

Prerequisites

Before running API_Testing.py:

  1. ✅ Temporal dev server running (temporal server start-dev)
  2. ✅ PostgreSQL database running
  3. ✅ Temporal worker running (python worker.py)
  4. ✅ REST API running (python run_api.py)

Check all services:

# Temporal
curl http://localhost:8233

# PostgreSQL
psql -d temporal_migration_test -c "SELECT 1"

# API
curl http://localhost:8000/health

# Worker
ps aux | grep "python worker.py"

Performance Comparison

Metric AWS Step Functions Temporal
Workflows started Sequential (1 at a time) Parallel (all at once)
Start time ~1 hour for 126 ~0.5 seconds for 124
Execution time ~1 hour total ~35 seconds total
Speedup Baseline ~103x faster

Troubleshooting

API Connection Errors

If you see connection refused errors:

# Check if API is running
curl http://localhost:8000/health

# If not, start it
cd temporal-implementation
python run_api.py

No Workers Available

If workflows don't complete:

# Check worker
ps aux | grep "python worker.py"

# Start if needed
cd temporal-implementation
python worker.py

Duplicate Name Errors

This is expected behavior. Duplicate names in names.txt will trigger the UPSERT mechanism, updating the existing record instead of failing.

Next Steps

With API_Testing.py now working with Temporal:

  1. ✅ All existing test workflows can use the new API
  2. ✅ Same test data and validation
  3. ✅ Same payload format (minimal code changes needed)
  4. ✅ Enhanced response data (workflow_id + run_id for tracking)
  5. ✅ Better performance (parallel execution)

The migration is complete and functionally equivalent to the AWS implementation!