AWS CloudWatch

Amazon CloudWatch provides monitoring and observability for AWS resources and applications. It collects metrics, logs, and events, enabling you to monitor, troubleshoot, and optimize your AWS environment.

Table of Contents

• Core Concepts
• Common Patterns
• CLI Reference
• Best Practices
• Troubleshooting
• References

Core Concepts

Metrics

Time-ordered data points published to CloudWatch. Key components:

• Namespace: Container for metrics (e.g., AWS/Lambda)
• Metric name: Name of the measurement (e.g., Invocations)
• Dimensions: Name-value pairs for filtering (e.g., FunctionName=MyFunc)
• Statistics: Aggregations (Sum, Average, Min, Max, SampleCount, pN)

Logs

Log data from AWS services and applications:

• Log groups: Collections of log streams
• Log streams: Sequences of log events from same source
• Log events: Individual log entries with timestamp and message

Alarms

Automated actions based on metric thresholds:

• States: OK, ALARM, INSUFFICIENT_DATA
• Actions: SNS notifications, Lambda functions, Auto Scaling, EC2 actions, Systems Manager OpsItems
• Types: metric (incl. metric math, anomaly detection, Metrics Insights), composite, log (scheduled Logs Insights query evaluated M-out-of-N on query results)
• Mute rules: scheduled windows that mute actions while alarms keep evaluating

Common Patterns

Create a Metric Alarm

AWS CLI:

# CPU utilization alarm for EC2
aws cloudwatch put-metric-alarm \
  --alarm-name "HighCPU-i-1234567890abcdef0" \
  --metric-name CPUUtilization \
  --namespace AWS/EC2 \
  --statistic Average \
  --period 300 \
  --threshold 80 \
  --comparison-operator GreaterThanThreshold \
  --evaluation-periods 2 \
  --dimensions Name=InstanceId,Value=i-1234567890abcdef0 \
  --alarm-actions arn:aws:sns:us-east-1:123456789012:alerts \
  --ok-actions arn:aws:sns:us-east-1:123456789012:alerts

boto3:

import boto3

cloudwatch = boto3.client('cloudwatch')

cloudwatch.put_metric_alarm(
    AlarmName='HighCPU-i-1234567890abcdef0',
    MetricName='CPUUtilization',
    Namespace='AWS/EC2',
    Statistic='Average',
    Period=300,
    Threshold=80.0,
    ComparisonOperator='GreaterThanThreshold',
    EvaluationPeriods=2,
    Dimensions=[
        {'Name': 'InstanceId', 'Value': 'i-1234567890abcdef0'}
    ],
    AlarmActions=['arn:aws:sns:us-east-1:123456789012:alerts'],
    OKActions=['arn:aws:sns:us-east-1:123456789012:alerts']
)

Lambda Error Rate Alarm

aws cloudwatch put-metric-alarm \
  --alarm-name "LambdaErrorRate-MyFunction" \
  --metrics '[
    {
      "Id": "errors",
      "MetricStat": {
        "Metric": {
          "Namespace": "AWS/Lambda",
          "MetricName": "Errors",
          "Dimensions": [{"Name": "FunctionName", "Value": "MyFunction"}]
        },
        "Period": 60,
        "Stat": "Sum"
      },
      "ReturnData": false
    },
    {
      "Id": "invocations",
      "MetricStat": {
        "Metric": {
          "Namespace": "AWS/Lambda",
          "MetricName": "Invocations",
          "Dimensions": [{"Name": "FunctionName", "Value": "MyFunction"}]
        },
        "Period": 60,
        "Stat": "Sum"
      },
      "ReturnData": false
    },
    {
      "Id": "errorRate",
      "Expression": "errors/invocations*100",
      "Label": "Error Rate",
      "ReturnData": true
    }
  ]' \
  --threshold 5 \
  --comparison-operator GreaterThanThreshold \
  --evaluation-periods 3 \
  --alarm-actions arn:aws:sns:us-east-1:123456789012:alerts

Query Logs with Insights

# Find errors in Lambda logs
aws logs start-query \
  --log-group-name /aws/lambda/MyFunction \
  --start-time $(date -d '1 hour ago' +%s) \
  --end-time $(date +%s) \
  --query-string '
    fields @timestamp, @message
    | filter @message like /ERROR/
    | sort @timestamp desc
    | limit 50
  '

# Get query results
aws logs get-query-results --query-id <query-id>

boto3:

import boto3
import time

logs = boto3.client('logs')

# Start query
response = logs.start_query(
    logGroupName='/aws/lambda/MyFunction',
    startTime=int(time.time()) - 3600,
    endTime=int(time.time()),
    queryString='''
        fields @timestamp, @message
        | filter @message like /ERROR/
        | sort @timestamp desc
        | limit 50
    '''
)

query_id = response['queryId']

# Wait for results
while True:
    result = logs.get_query_results(queryId=query_id)
    if result['status'] == 'Complete':
        break
    time.sleep(1)

for row in result['results']:
    print(row)

Create Metric Filter

Extract metrics from log patterns:

# Create metric filter for error count
aws logs put-metric-filter \
  --log-group-name /aws/lambda/MyFunction \
  --filter-name ErrorCount \
  --filter-pattern "ERROR" \
  --metric-transformations \
    metricName=ErrorCount,metricNamespace=MyApp,metricValue=1,defaultValue=0

Create a Log Alarm

Alarm directly on a Logs Insights query (no metric filter needed). CloudWatch creates and manages the underlying scheduled query. The ScheduledQueryRoleARN role must trust logs.amazonaws.com and allow logs:StartQuery and logs:GetQueryResults on the log group ARNs, plus logs:StopQuery and logs:DescribeLogGroups on "Resource": "*" (these two support no resource-level permissions, so a log-group-scoped statement never matches them).

# ALARM when >100 errors in 3 of the last 5 query runs
aws cloudwatch put-log-alarm \
  --alarm-name "HighErrorCount" \
  --comparison-operator GreaterThanThreshold \
  --threshold 100 \
  --query-results-to-evaluate 5 \
  --query-results-to-alarm 3 \
  --treat-missing-data notBreaching \
  --alarm-actions arn:aws:sns:us-east-1:123456789012:alerts \
  --scheduled-query-configuration '{
    "QueryString": "fields @timestamp, @message | filter @message like /ERROR/",
    "LogGroupIdentifiers": ["/aws/lambda/MyFunction"],
    "ScheduledQueryRoleARN": "arn:aws:iam::123456789012:role/ScheduledQueryRole",
    "AggregationExpression": "count(*)",
    "ScheduleConfiguration": {
      "ScheduleExpression": "rate(10 minutes)",
      "StartTimeOffset": 600
    }
  }'

# Log alarms are omitted from describe-alarms unless requested
aws cloudwatch describe-alarms --alarm-types LogAlarm

Publish Custom Metrics

import boto3

cloudwatch = boto3.client('cloudwatch')

cloudwatch.put_metric_data(
    Namespace='MyApp',
    MetricData=[
        {
            'MetricName': 'OrdersProcessed',
            'Value': 1,
            'Unit': 'Count',
            'Dimensions': [
                {'Name': 'Environment', 'Value': 'Production'},
                {'Name': 'OrderType', 'Value': 'Standard'}
            ]
        }
    ]
)

Create Dashboard

cat > dashboard.json << 'EOF'
{
  "widgets": [
    {
      "type": "metric",
      "x": 0, "y": 0, "width": 12, "height": 6,
      "properties": {
        "title": "Lambda Invocations",
        "metrics": [
          ["AWS/Lambda", "Invocations", "FunctionName", "MyFunction"]
        ],
        "period": 60,
        "stat": "Sum",
        "region": "us-east-1"
      }
    },
    {
      "type": "log",
      "x": 12, "y": 0, "width": 12, "height": 6,
      "properties": {
        "title": "Recent Errors",
        "query": "SOURCE '/aws/lambda/MyFunction' | filter @message like /ERROR/ | limit 20",
        "region": "us-east-1"
      }
    }
  ]
}
EOF

aws cloudwatch put-dashboard \
  --dashboard-name MyAppDashboard \
  --dashboard-body file://dashboard.json

CLI Reference

Metrics Commands

Alarms Commands

Logs Commands

Best Practices

Metrics

• Use dimensions wisely — too many creates metric explosion
• Aggregate before publishing — batch custom metrics
• Use high-resolution metrics (1-second) only when needed
• Set meaningful units for custom metrics

Alarms

• Use composite alarms for complex conditions
• Set appropriate evaluation periods to avoid flapping
• Include OK actions to track recovery
• Use anomaly detection for dynamic thresholds
• Use log alarms instead of metric filter + metric alarm for query-based conditions; set --treat-missing-data notBreaching for sparse errors, breaching to detect logs that stop arriving
• Stagger log alarm schedules — concurrent scheduled query executions per account are capped at 100
• Set --warm-up-configuration on alarms created alongside new resources (1-2880 min) to avoid noise before metrics publish
• Use --evaluation-window WallClockWindow={Timezone=...} for daily/weekly batch or backup alarms; keep the default sliding window for Auto Scaling
• Use mute rules for maintenance, not disable-alarm-actions; enable-alarm-actions does not unmute an active mute rule

Logs

• Set retention policies — don't keep logs forever
• Use structured logging (JSON) for better querying
• Create metric filters for key events
• Use Contributor Insights for top-N analysis

Cost Optimization

• Delete unused dashboards
• Reduce log retention for non-critical logs
• Avoid high-resolution metrics unless necessary
• Use log subscription filters instead of polling

Troubleshooting

Missing Metrics

Causes:

• Service not publishing yet (wait 1-5 minutes)
• Wrong namespace/dimensions
• Detailed monitoring not enabled (EC2)

Debug:

# List metrics for a namespace
aws cloudwatch list-metrics \
  --namespace AWS/Lambda \
  --dimensions Name=FunctionName,Value=MyFunction

Alarm Stuck in INSUFFICIENT_DATA

Causes:

• Metric not being published
• Dimensions mismatch
• Evaluation period too short
• Warm-up period still active (ends early once data fills the window unless OnlyStartEvaluatingAfterWarmUpPeriodEnds=true)
• Log alarm: just created/updated (query, schedule, or log group changes reset state), scheduled query role lacks permissions (EvaluationState = EVALUATION_ERROR, see StateReason), or ingestion lag (shift window back with EndTimeOffset)

Debug:

# Check if metric has data
aws cloudwatch get-metric-statistics \
  --namespace AWS/Lambda \
  --metric-name Invocations \
  --dimensions Name=FunctionName,Value=MyFunction \
  --start-time $(date -d '1 hour ago' -u +%Y-%m-%dT%H:%M:%SZ) \
  --end-time $(date -u +%Y-%m-%dT%H:%M:%SZ) \
  --period 60 \
  --statistics Sum

# Log alarm: check evaluation state and reason
aws cloudwatch describe-alarms --alarm-types LogAlarm --alarm-names HighErrorCount \
  --query 'LogAlarms[].[StateValue,EvaluationState,StateReason]'

Log Events Not Appearing

Causes:

• IAM permissions missing
• CloudWatch Logs agent not running
• Log group doesn't exist

Debug:

# Check log streams
aws logs describe-log-streams \
  --log-group-name /aws/lambda/MyFunction \
  --order-by LastEventTime \
  --descending \
  --limit 5

High CloudWatch Costs

Check usage:

# Get PutLogEvents usage
aws cloudwatch get-metric-statistics \
  --namespace AWS/Logs \
  --metric-name IncomingBytes \
  --dimensions Name=LogGroupName,Value=/aws/lambda/MyFunction \
  --start-time $(date -d '7 days ago' -u +%Y-%m-%dT%H:%M:%SZ) \
  --end-time $(date -u +%Y-%m-%dT%H:%M:%SZ) \
  --period 86400 \
  --statistics Sum

References

• CloudWatch User Guide
• CloudWatch Logs User Guide
• CloudWatch API Reference
• CloudWatch CLI Reference
• Logs Insights Query Syntax
• boto3 CloudWatch