نظارت فعال عملکرد API: اطمینان از به روزرسانی

Payment API failure for an e-commerce site:
- ~$10k in lost revenue per hour
- Abandoned carts
- Frustrated customers
- Social media complaints
- Urgent all-hands incident response

// Response time distribution can tell you more than averages
const responseTimeBuckets = {
  "0-100ms": 65,    // 65% of requests
  "100-300ms": 25,  // 25% of requests
  "300-500ms": 7,   // 7% of requests
  "500ms+": 3       // 3% of requests (investigate these!)
};

// Breaking down errors by type is more useful than overall rates
const errorBreakdown = {
  "5xx": 37,      // Server errors - your fault
  "4xx": 158,     // Client errors - could be your fault
  "Timeouts": 42, // Connection issues - investigate
  "Auth": 89      // Auth failures - check token management
};

# Example throughput monitoring query
$ curl -s https://api.metrics.example.com/v1/throughput/last-hour | jq
{
  "total_requests": 145782,
  "avg_rps": 40.5,
  "peak_rps": 178.3,
  "peak_time": "2023-02-15T12:34:21Z"
}

# Availability check command
$ uptime -d https://api.example.com/health
Endpoint: https://api.example.com/health
Status: UP
Uptime: 99.97% (Last 30 days)
Outages: 1 (Total duration: 12m 34s)
Last outage: 2023-02-10T03:15:22Z to 2023-02-10T03:27:56Z

# Basic health check (surface level)
$ curl -s https://api.example.com/health
{"status":"UP","version":"2.3.1"}

# Deeper synthetic transaction (functional check)
$ curl -s -X POST https://api.example.com/v1/orders \
  -H "Authorization: Bearer $TOKEN" \
  -d '{"product_id":"test-123","quantity":1}' | jq
{
  "success": true,
  "order_id": "ord_test_7f3a5",
  "status": "created"
}

// Pseudocode for circuit breaker pattern
function callDependencyAPI(request) {
  if (circuitBreaker.isOpen()) {
    return fallbackResponse(); // Don't even try if circuit is open
  }

  try {
    const response = await sendRequest(request);
    circuitBreaker.recordSuccess();
    return response;
  } catch (error) {
    circuitBreaker.recordFailure();
    return fallbackResponse();
  }
}

API Response Time Spike at 14:32:15
↓
Database CPU Usage Spike at 14:32:10
↓
Backup Job Started at 14:30:00

# Set up monitoring from multiple regions
for region in us-east eu-west ap-south; do
  monitor create \
    --name "api-health-$region" \
    --url "https://api.example.com/health" \
    --region $region \
    --interval 30s \
    --alert-threshold 5s
done

API Service
├── Container Metrics
│   ├── CPU Usage
│   ├── Memory Usage
│   └── Network I/O
├── JVM Metrics (if applicable)
│   ├── Heap Usage
│   ├── Garbage Collection
│   └── Thread Count
└── Dependencies
    ├── Database Connection Pool
    ├── Cache Hit Rate
    └── External Service Response Times

// Example correlation between API errors and cart abandonment
const apiErrorRates = [2.1, 3.5, 7.8, 12.4, 4.2, 2.8];
const cartAbandonment = [3.2, 4.1, 8.5, 15.2, 6.1, 3.5];

// Correlation shows clear relationship between these metrics
const correlation = calculateCorrelation(apiErrorRates, cartAbandonment);
console.log(`Correlation coefficient: ${correlation.toFixed(2)}`); // 0.97

API Service Level Objectives:
- Availability: 99.95% uptime (21.9 minutes downtime/month maximum)
- Latency: 95% of requests complete in < 200ms
- Error Rate: < 0.1% of requests result in 5xx errors

# Create a simple uptime monitor with curl
while true; do
  start_time=$(date +%s.%N)
  http_status=$(curl -s -o /dev/null -w "%{http_code}" https://api.example.com/health)
  end_time=$(date +%s.%N)
  latency=$(echo "$end_time - $start_time" | bc)

  if [[ $http_status -ne 200 ]]; then
    echo "$(date) - API health check failed: $http_status"
    # Send alert via webhook, email, etc.
  fi

  echo "$(date) - Status: $http_status, Latency: ${latency}s"
  sleep 60
done

1. Set up synthetic transactions for key user flows
2. Implement dependency monitoring
3. Create dashboards that visualize API health
4. Configure alerting with appropriate thresholds
5. Establish on-call procedures for incident response

1. Weekly review of monitoring data
2. Identify patterns and trends
3. Set performance improvement goals
4. Implement changes
5. Measure impact

Critical APIs monitored:
- Product catalog API
- Search API
- Cart/checkout API
- Payment processing API
- User authentication API

Monitoring approach:
1. Health checks every 30 seconds
2. Synthetic transactions every 5 minutes
3. Response time thresholds based on 95th percentile
4. Separate monitoring for mobile vs. web API endpoints
5. Alerts routed to appropriate teams based on component