AWS for DevOps — X-Ray & SNS Notifications

10.5 — X-Ray & SNS Notifications

X-Ray traces requests across distributed services. SNS delivers alerts to your team. Together, they close the loop: detect → understand → notify.

AWS X-Ray

How X-Ray Works

  ┌──────┐     ┌──────────┐     ┌──────────┐     ┌─────────┐
  │Client│────→│API Gateway│────→│Lambda Fn │────→│DynamoDB  │
  └──────┘     └──────────┘     └──────────┘     └─────────┘
                    │                 │                 │
                    ▼                 ▼                 ▼
               ┌─────────────────────────────────────────┐
               │         X-Ray — Trace Map               │
               │  Shows latency, errors, connections     │
               └─────────────────────────────────────────┘

Real-World Analogy

X-Ray is like a package tracking system. When you order something online, you can see: shipped from warehouse → arrived at sorting center → out for delivery → delivered. X-Ray does the same for API requests through your microservices.

Enabling X-Ray on Lambda

# Update a Lambda function to enable X-Ray tracing
$ aws lambda update-function-configuration \
    --function-name my-api-handler \
    --tracing-config Mode=Active
{
    "FunctionName": "my-api-handler",
    "TracingConfig": {
        "Mode": "Active"
    }
}

Querying Traces

# Get trace summaries for the last 10 minutes
$ aws xray get-trace-summaries \
    --start-time $(date -u -v-10M +%s) \
    --end-time $(date -u +%s) \
    --filter-expression 'service("my-api-handler") AND responsetime > 1'
{
    "TraceSummaries": [
        {
            "Id": "1-abc123-def456",
            "Duration": 2.3,
            "ResponseTime": 2.1,
            "Http": {
                "HttpURL": "https://api.example.com/users",
                "HttpStatus": 200,
                "HttpMethod": "GET"
            }
        }
    ]
}

# Get full trace details
$ aws xray batch-get-traces --trace-ids "1-abc123-def456"

X-Ray Service Map

The service map shows how your services are connected and where bottlenecks exist:

Metric	What It Shows	Action
Response time	How long each service takes	Optimize slow services
Error rate	Percentage of failed requests	Fix failing services
Throughput	Requests per second	Scale busy services
Connections	Which services call which	Identify coupling

Amazon SNS (Simple Notification Service)

SNS is a pub/sub messaging service. For DevOps, it’s the primary way to deliver alerts from CloudWatch Alarms to your team.

Creating an SNS Topic and Subscription

# Create a topic for alerts
$ aws sns create-topic --name devops-alerts
{
    "TopicArn": "arn:aws:sns:ap-south-1:123456789012:devops-alerts"
}

# Subscribe an email
$ aws sns subscribe \
    --topic-arn arn:aws:sns:ap-south-1:123456789012:devops-alerts \
    --protocol email \
    --notification-endpoint oncall@example.com
{
    "SubscriptionArn": "pending confirmation"
}

# Subscribe a Slack webhook via HTTPS
$ aws sns subscribe \
    --topic-arn arn:aws:sns:ap-south-1:123456789012:devops-alerts \
    --protocol https \
    --notification-endpoint https://hooks.slack.com/services/T00/B00/xxx

Connecting CloudWatch Alarms to SNS

# Create an alarm that notifies via SNS
$ aws cloudwatch put-metric-alarm \
    --alarm-name high-error-rate \
    --metric-name 5XXError \
    --namespace AWS/ApiGateway \
    --statistic Sum \
    --period 300 \
    --threshold 10 \
    --comparison-operator GreaterThanThreshold \
    --evaluation-periods 2 \
    --alarm-actions arn:aws:sns:ap-south-1:123456789012:devops-alerts \
    --ok-actions arn:aws:sns:ap-south-1:123456789012:devops-alerts

Notification Flow

  CloudWatch Alarm  ──→  SNS Topic  ──→  Email
  (5XX > 10)             │               Slack (via HTTPS)
                         │               PagerDuty (via HTTPS)
                         └──→  Lambda ──→ Custom processing

Alert Routing Strategy

Alert Level	SNS Topic	Subscribers	Example
P1 — Critical	critical-alerts	PagerDuty + phone	Service down, data loss
P2 — High	devops-alerts	Slack #incidents + email	High error rate, latency spike
P3 — Warning	warning-alerts	Slack #monitoring	Disk 80%, approaching limits
P4 — Info	info-alerts	Slack #deployments	Deploy complete, scaling event

Common Mistakes

Mistake	Why It Hurts	Fix
One SNS topic for everything	Alert fatigue — all alerts look the same	Separate topics by severity
No OK action on alarms	Team doesn’t know when issue resolves	Add --ok-actions to alarms
Email-only notifications	Slow response — who checks email at 3 AM?	Use PagerDuty/Slack for critical alerts
Not enabling X-Ray	Can’t trace issues across services	Enable Active tracing on Lambda/API Gateway

Quick Reference

Task	Command
Create SNS topic	aws sns create-topic --name NAME
Subscribe email	aws sns subscribe --topic-arn ARN --protocol email --notification-endpoint EMAIL
Publish test message	aws sns publish --topic-arn ARN --subject "Test" --message "Hello"
List subscriptions	aws sns list-subscriptions-by-topic --topic-arn ARN
Enable X-Ray on Lambda	aws lambda update-function-configuration --function-name FN --tracing-config Mode=Active
Get X-Ray traces	aws xray get-trace-summaries --start-time T1 --end-time T2

---

Next → Monitoring Exercises