email notifications

Lambda function that provide email notifications

To run the lambda function, you need to create a task in the Galloper and specify a lambda handler in it.

lambda_function.email_lambda_handler - handler for API tests notifications

lambda_function.ui_email_lambda_handler - handler for UI tests notifications

You can use curl to invoke a task, example below

curl -XPOST -H "Content-Type: application/json"
    -d '{"param1": "value1", "param2": "value2", ...}' <host>:5000/task/<task_id>

<host> - Galloper host DNS or IP

<task_id> - ID of the created task in Galloper

You can pass the necessary parameters with the -d option. List of available parameters:

'test': '<ui_scenario_name>' - required for ui email notifications

'test_suite': '<ui_suite_name>' - required for ui email notifications

'test': '<simulation_name>' - required for api email notifications

'test_type': '<test_type>' - required for api email notifications

'users': '<count_of_vUsers>' - required for all type of notifications

'influx_host': '<influx_host_DNS_or_IP>' - required for all type of notifications

'smpt_user': '<smpt_user_who_will_login_to_the_the_host>' - required for all type of notifications - note: parameter name smpt instead of smtp

'smpt_password': '<password>' - required for all type of notifications - note: parameter name smpt instead of smtp

'smpt_sender': '<sender_who_email_will_be_FROM>' - optional: if not included then 'smpt_user' will be used - note: parameter name smpt instead of smtp

'user_list': '<list of recipients>' - required for all type of notifications

'notification_type': '<test_type>' - should be 'ui' or 'api'

'smpt_host': 'smtp.gmail.com' - optional, default - 'smtp.gmail.com' - note: parameter name smpt instead of smtp

'smpt_port': 465 - optional, default - 465 - note: parameter name smpt instead of smtp

'influx_port': 8086 - optional, default - 8086

'influx_thresholds_database': 'thresholds' - optional, default - 'thresholds'

'influx_ui_tests_database': 'perfui' - optional, default - 'perfui'

'influx_comparison_database': 'comparison' - optional, default - 'comparison'

'influx_user': '' - optional, default - ''

'influx_password': '' - optional, default - ''

'test_limit': 5 - optional, default - 5

'comparison_metric': 'pct95' - optional, only for api notifications, default - 'pct95'

---

AI-Powered Performance Analysis (Backend Notifications)

Backend (API) email notifications support AI-powered performance analysis using Azure OpenAI. When enabled, emails include an "AI Analysis" section with automated insights based on test results and SLA thresholds.

Features

• Key Findings: 3-5 bullet points highlighting critical performance issues
• Performance Observations: Markdown table with problematic transactions/requests
• Actionable Items: 3-5 numbered, data-specific recommendations

AI Analysis Parameters

All AI parameters are optional and backward-compatible. Emails send successfully even if AI analysis fails (graceful degradation).

'enable_ai_analysis': true/false - optional, default: false - Enable AI-powered analysis section

'ai_provider': 'azure_openai' - optional, default: 'azure_openai' - AI provider type (currently only Azure OpenAI supported)

'azure_openai_api_key': '<your-api-key>' - required if AI enabled - Azure OpenAI API key for authentication

'azure_openai_endpoint': 'https://your-resource.openai.azure.com/' - required if AI enabled - Azure OpenAI endpoint URL

'azure_openai_api_version': '2024-02-15-preview' - optional, default: '2024-02-15-preview' - API version

'ai_model': 'gpt-4o' - optional, default: 'gpt-4o' - Model deployment name (must match Azure deployment)

'ai_temperature': 0.0 - optional, default: 0.0 - Temperature for LLM output (0.0 = deterministic, 2.0 = creative)

Example Usage

curl -XPOST -H "Content-Type: application/json" -d '{
  "test": "checkout_flow_test",
  "test_type": "backend",
  "notification_type": "api",
  "enable_ai_analysis": true,
  "azure_openai_api_key": "sk-...",
  "azure_openai_endpoint": "https://my-resource.openai.azure.com/",
  "ai_model": "gpt-4o",
  "influx_host": "influxdb.example.com",
  "smtp_user": "notifications@example.com",
  "smtp_password": "password",
  "user_list": "team@example.com"
}' <host>:5000/task/<task_id>

Graceful Degradation

If AI analysis fails (invalid credentials, timeout, API errors), the email will still send successfully without the AI Analysis section. This ensures core reporting functionality is never blocked by AI service issues.

Requirements

• Azure OpenAI resource with GPT-4o model deployment
• API key with appropriate permissions
• Lambda timeout configured to at least 120 seconds (AI generation adds ~10s per email)

Provider Extensibility

The implementation uses provider abstraction to support future AI providers (OpenAI, Anthropic Claude, etc.) without code changes. See specs/002-backend-ai-analysis/quickstart.md for details on adding new providers.