Deep Dive: MCP Integration

HelpMeTest

Installation
Available MCP Tools
Test Management
Health Check Monitoring
Advanced Usage Patterns
Real-World Examples
Troubleshooting

Installation

Prerequisites

HelpMeTest API token (get it from your account settings)
One of the supported AI editors:
- Claude Desktop
- Cursor
- VSCode with Claude extension
- Any MCP-compatible editor

Quick Installation

The installer will:

Detect which AI editors you have installed
Configure MCP integration for each detected editor
Provide setup instructions or open configuration links
Verify the installation

Manual Configuration

If automatic detection doesn't work or you prefer manual setup:

For Claude Desktop:

Add to ~/Library/Application Support/Claude/claude_desktop_config.json:

For Cursor:

Add to your Cursor MCP settings (Settings → MCP → Add Server):

After configuration, restart your AI editor for changes to take effect.

Available MCP Tools

Test Management Tools

`helpmetest_list_tests`

Discovers and lists all available tests in your account.

What your AI can do:

List all tests with names, descriptions, and tags
Organize tests by category (uptime, API, user journeys, etc.)
Filter tests by tags
Show test counts and organization
Recommend which tests to run for specific scenarios

Example interactions:

You: "What tests do I have?"
AI: "You have 10 tests organized by category:
    🌐 Uptime Tests (tag: uptime) - 4 tests
    • All your main domains and services
    🔍 Functional Tests (tag: keywords) - 2 tests
    • Search functionality, form submissions
    👤 User Journey Tests (tag: cuj) - 2 tests
    • Login flow, critical user paths
    📄 Content Tests (tag: docs, landing) - 2 tests
    • Documentation, landing pages"

`helpmetest_run_test`

Executes tests by name, ID, or tag and returns detailed results.

Parameters:

identifier: Test name, ID, or tag (with tag: prefix)

What your AI can do:

Execute individual tests by name
Run all tests with a specific tag
Execute tests by ID
Analyze test results and performance
Identify failures and suggest debugging steps
Compare results against historical data

Example interactions:

Run by name:

You: "Execute the Google search test"
AI: "Running 'Google search' test...
    ❌ Test FAILED
    • Duration: 2.3 seconds
    • Issue: Search automation blocked
    • Cause: Rate limiting or CAPTCHA protection
    💡 Analysis:
    This is a common issue with automated search testing.
    The test logic is working, but Google's anti-bot measures
    are preventing automation. Consider updating the test
    to use a different search endpoint or mock the search."

Run by tag:

You: "Run all uptime tests"
AI: "Executing uptime test suite...
    ✅ All 4 uptime tests PASSED!
    📊 Results:
    • netdata.helpmetest.com: ✅ PASS (185ms)
    • login.helpmetest.com: ✅ PASS (127ms)
    • ntfy.helpmetest.com: ✅ PASS (20ms)
    • helpmetest.com: ✅ PASS (22ms)
    🎯 Analysis:
    • Total execution: 0.35 seconds
    • Success rate: 100%
    • All response times normal
    • No performance issues detected"

Run by ID:

Health Check Tools

`helpmetest_status_health`

Queries health check status for all configured services.

What your AI can do:

List all health checks with current status
Show last check time and grace periods
Identify unhealthy services
Explain what each health check monitors
Suggest actions for failing checks

Example interactions:

You: "Show me health check status"
AI: "Current health check status:
    ✅ web-frontend (last check: 2 minutes ago)
    • Grace period: 5 minutes
    • Status: Healthy
    • Last heartbeat: HTTP 200 OK
    ✅ api-backend (last check: 1 minute ago)
    • Grace period: 2 minutes
    • Status: Healthy
    • Last heartbeat: API responding normally
    ⚠️  background-worker (last check: 8 minutes ago)
    • Grace period: 10 minutes
    • Status: Warning - approaching grace period
    • Last heartbeat: 8 minutes ago
    Recommendation: Check background-worker - it's close to
    exceeding its grace period."

Test Management

Test Organization with Tags

Tags help organize and filter tests for different scenarios:

Common tag patterns:

uptime - Service availability and connectivity tests
api - API endpoint and backend service tests
cuj - Critical user journey and workflow tests
keywords - Search and functionality tests
docs - Documentation and content tests
landing - Landing page and marketing tests
flaky - Tests that may be unstable

Using tags effectively:

You: "What tests should I run before deploying a search feature update?"
AI: "For a search feature deployment, I recommend:
    🔍 Search-Related Tests (keywords tag):
    • Google search test
    • Landing page search functionality
    🌐 Core Service Tests (uptime tag):
    • Main application uptime
    • API endpoint health
    👤 User Impact Tests (cuj tag):
    • Landing -> /docs navigation (affects search UX)
    Would you like me to run this recommended test suite?"

Pre-Deployment Testing Workflows

Complete pre-deployment check:

You: "Before I deploy, verify everything is ready"
AI: "Running comprehensive pre-deployment check...
    📋 Step 1: Service Health
    ✅ All 3 health checks healthy
    • web-frontend, api-backend, background-worker all reporting
    📋 Step 2: Critical Tests
    Running uptime tests... ✅ 4/4 PASSED
    Running API tests... ✅ 3/3 PASSED
    Running user journey tests... ✅ 2/2 PASSED
    📋 Step 3: Performance Check
    • Average response time: 156ms (normal)
    • All services within acceptable ranges
    • No performance degradation detected
    🎉 Deployment Readiness: APPROVED
    All systems operational, no blocking issues found."

Deployment verification:

You: "Did the deployment go well? Run post-deployment verification"
AI: "Running post-deployment verification...
    ✅ Uptime Tests: 4/4 PASSED
    • All main services responding normally
    ✅ API Health Tests: 3/3 PASSED
    • All endpoints returning expected responses
    ✅ Critical User Journeys: 2/2 PASSED
    • Login flow working
    • Core functionality verified
    🎉 Deployment verification complete - all systems operational!"

Health Check Monitoring

Understanding Health Check Status

Health checks report service status through heartbeat signals:

Status indicators:

✅ Healthy: Last heartbeat within grace period
⚠️ Warning: Approaching grace period limit
❌ Down: Exceeded grace period, no recent heartbeat

Grace periods define how long a service can go without reporting before being marked as down. They vary by service type:

Web services: 1-5 minutes
API services: 30 seconds - 2 minutes
Background workers: 10-15 minutes
Batch jobs: Hours to days depending on schedule

Service Health Queries

You: "Are all our critical services responding?"
AI: "Checking critical service health...
    ✅ Main API (api.myapp.com): Healthy
    • Last check: 45 seconds ago
    • Response time: 156ms
    • Status: All systems operational
    ✅ Auth service (auth.myapp.com): Healthy
    • Last check: 1 minute ago
    • Response time: 89ms
    • Status: Normal operation
    ✅ CDN (cdn.myapp.com): Healthy
    • Last check: 30 seconds ago
    • Response time: 23ms
    • Status: Excellent performance
    All critical services are healthy and responding within normal parameters."

Advanced Usage Patterns

Performance Analysis

You: "Are my tests running slower than usual?"
AI: "Analyzing test performance trends...
    📈 Performance Analysis:
    • Current uptime tests: 0.35s (normal)
    • Historical average: 0.28s
    • Trend: 25% slower than baseline
    🔍 Breakdown:
    • netdata.helpmetest.com: 185ms (was 120ms avg)
    • Other services: Normal response times
    💡 Recommendation:
    The netdata service is responding slower than usual.
    This could indicate increased load or performance issues.
    Consider investigating the monitoring dashboard performance."

Failure Analysis and Debugging

You: "My login test is failing. Help me debug it."
AI: "Analyzing login test failure...
    🔍 Failure Analysis:
    • Test: Login user journey
    • Failure point: Authentication step
    • Error: Timeout waiting for redirect
    • Duration: 30s (exceeded timeout)
    🧐 Possible Causes:
    1. Authentication service slow/down
    2. Frontend JavaScript errors
    3. Network connectivity issues
    4. Test environment configuration
    🔧 Debugging Steps:
    Let me check the auth service health first...
    Auth service status: ⚠️  Responding slowly (2.3s vs normal 89ms)
    This explains your login test timeout.
    Recommendation: Investigate auth service performance.
    The slow response is causing your test to timeout."

Intelligent Test Recommendations

You: "What tests should I run for a database migration?"
AI: "For a database migration, I recommend this test strategy:
    🔴 Critical Priority (run first):
    • API tests (tag:api) - Verify database connections
    • Login test - Check authentication still works
    🟡 Medium Priority (after migration):
    • User journey tests (tag:cuj) - Verify data integrity
    • Uptime tests - Confirm all services recovered
    🟢 Optional (if time permits):
    • Performance tests - Check for query slowdowns
    • Content tests - Verify data display
    Would you like me to run the critical tests first?"

Real-Time Execution Monitoring

You: "Run the user journey tests and show me detailed progress"
AI: "Starting user journey test execution...
    🚀 Test 1/2: Login flow
    • Initializing browser... ✅
    • Loading login page... ✅ (89ms)
    • Entering credentials... ✅
    • Submitting form... ✅
    • Verifying redirect... ✅
    Result: ✅ PASSED (2.3s)
    🚀 Test 2/2: Landing -> /docs
    • Loading landing page... ✅ (156ms)
    • Finding docs link... ✅
    • Clicking navigation... ✅
    • Verifying docs page... ✅
    Result: ✅ PASSED (1.8s)
    📊 Summary: 2/2 tests passed, total time 4.1s"

Real-World Examples

For Developers

Development workflow:

Before demos:

For DevOps Teams

Daily standup prep:

You: "Give me a quick status of all our services for standup"
AI: "Service status overview for standup:
    ✅ Production Services: All healthy
    • 4/4 uptime checks reporting normally
    • Average response time: 145ms
    ⚠️  Background Jobs: One warning
    • data-processor: Last heartbeat 8 min ago (10 min grace period)
    📊 Last 24 Hours:
    • 0 incidents
    • 100% uptime
    • All tests passing
    Note: Monitor data-processor - approaching grace period limit."

Deployment verification:

You: "Did the deployment go well? Run the post-deployment test suite"
AI: "Running post-deployment verification tests...
    ✅ Uptime Tests: 4/4 PASSED
    ✅ API Health Tests: 3/3 PASSED
    ✅ Critical User Journeys: 2/2 PASSED
    🎉 Deployment verification complete!
    • All services responding normally
    • No errors detected
    • Performance within acceptable ranges
    Deployment was successful."

For SRE Teams

Incident response:

You: "I got an alert about the API, can you check if it's responding?"
AI: "Checking API health immediately...
    ❌ Issue confirmed: API is returning 500 errors
    • URL: https://api.myapp.com/health
    • Response: 500 Internal Server Error
    • Response time: 2.3 seconds
    This suggests an application-level problem rather than
    complete outage. The server is responding but returning errors.
    Recommendation: Check application logs and database connectivity."

Performance investigation:

You: "How are our endpoints performing right now?"
AI: "Current performance status:
    ✅ Main API (api.myapp.com): 156ms - Good
    ⚠️  Auth service (auth.myapp.com): 892ms - Slow
    ✅ CDN (cdn.myapp.com): 23ms - Excellent
    Alert: Auth service is 10x slower than normal (89ms baseline).
    This will impact login performance and user experience.
    Recommendation: Investigate auth service immediately."

For QA Teams

Test environment validation:

Test execution organization:

You: "Run all tests related to the payment flow"
AI: "Searching for payment-related tests...
    Found 3 payment tests:
    • Payment form submission (tag:payments, cuj)
    • Payment API integration (tag:api, payments)
    • Payment confirmation (tag:payments)
    Running all payment tests...
    ✅ Payment form submission: PASSED (3.2s)
    ✅ Payment API integration: PASSED (1.8s)
    ✅ Payment confirmation: PASSED (2.1s)
    All payment functionality verified and working correctly."

Troubleshooting

No Tests Found

Symptom: Your AI reports "You currently have no tests configured"

Causes:

No tests have been created in your account yet
API token doesn't have access to test data
Tests exist but haven't synchronized yet

Solutions:

Verify you have tests created:
- Check the HelpMeTest web interface
- Ensure tests are saved and not just drafts
Check API token permissions:
- Regenerate your API token if needed
- Update MCP configuration with new token
- Restart your AI editor
Create a test to verify:

MCP Server Not Responding

Symptom: AI says "I can't access HelpMeTest tools right now"

Causes:

MCP server not configured properly
CLI not in system PATH
API token missing or invalid

Solutions:

Verify CLI installation:
Test MCP server directly:
Should start the MCP server without errors.
Check MCP configuration:
- Verify API token in config file
- Ensure helpmetest command is accessible
- Restart your AI editor after configuration changes

Test Execution Failures

Symptom: Tests fail unexpectedly or consistently

Debugging steps:

Ask AI to analyze the failure:
Check service dependencies:
Verify test environment:

Performance Issues

Symptom: Tests running slower than expected

Analysis:

You: "Why are my tests running slow?"
AI: "Analyzing test performance...
    Current execution time: 4.2 seconds
    Historical average: 2.8 seconds
    Slowdown: 50% slower than baseline
    Breakdown:
    • API test: 2.1s (was 1.2s) - 75% slower
    • Login test: 2.1s (was 1.6s) - 31% slower
    The API test is the primary bottleneck.
    Recommendation: Check API server performance and load."

Health Check Configuration Issues

Symptom: Health checks showing incorrect status

Common issues:

Grace period too short
- Services marked down when they're actually healthy
- Solution: Increase grace period to match service characteristics
Grace period too long
- Failures not detected quickly enough
- Solution: Decrease grace period for faster detection
Heartbeat not reporting
- Health checks never update
- Solution: Verify the service is calling helpmetest health command

Ask your AI for help:

You: "My background worker health check is always showing as down but the service is running"
AI: "Let me check the health check configuration...
    Current grace period: 5 minutes
    Last heartbeat: 8 minutes ago
    Issue: The service exceeded the grace period. This means either:
    1. The service isn't sending heartbeats (most likely)
    2. The grace period is too short for this service type
    Background workers typically need 10-15 minute grace periods.
    Recommendation: Update your health check to use a longer grace period:
    helpmetest health 'background-worker' '15m'"

This completes the technical deep dive. Your AI assistant now has comprehensive capabilities for testing and monitoring. For the high-level overview and quick start guide, see the main MCP Integration guide.

Questions? Email us at contact@helpmetest.com

Table of Contents

Installation

Prerequisites

Quick Installation

Manual Configuration

Available MCP Tools

Test Management Tools

helpmetest_list_tests

helpmetest_run_test

Health Check Tools

helpmetest_status_health

Test Management

Test Organization with Tags

Pre-Deployment Testing Workflows

Health Check Monitoring

Understanding Health Check Status

Service Health Queries

Advanced Usage Patterns

Performance Analysis

Failure Analysis and Debugging

Intelligent Test Recommendations

Real-Time Execution Monitoring

Real-World Examples

For Developers

For DevOps Teams

For SRE Teams

For QA Teams

Troubleshooting

No Tests Found

MCP Server Not Responding

Test Execution Failures

Performance Issues

Health Check Configuration Issues

`helpmetest_list_tests`

`helpmetest_run_test`

`helpmetest_status_health`