AWS for DevOps — CloudWatch Dashboards

10.3 — CloudWatch Dashboards

Dashboards give you a single-pane view of your system’s health. Build them for your team, your on-call, and your stakeholders.

Creating a Dashboard

$ aws cloudwatch put-dashboard \
    --dashboard-name production-overview \
    --dashboard-body '{
        "widgets": [
            {
                "type": "metric",
                "x": 0, "y": 0, "width": 12, "height": 6,
                "properties": {
                    "title": "ECS CPU Utilization",
                    "metrics": [
                        ["AWS/ECS", "CPUUtilization", "ServiceName", "api", "ClusterName", "production"]
                    ],
                    "period": 300,
                    "stat": "Average",
                    "region": "ap-south-1"
                }
            },
            {
                "type": "metric",
                "x": 12, "y": 0, "width": 12, "height": 6,
                "properties": {
                    "title": "ALB 5XX Errors",
                    "metrics": [
                        ["AWS/ApplicationELB", "HTTPCode_Target_5XX_Count", "LoadBalancer", "app/my-alb/abc123"]
                    ],
                    "period": 60,
                    "stat": "Sum",
                    "region": "ap-south-1"
                }
            },
            {
                "type": "log",
                "x": 0, "y": 6, "width": 24, "height": 6,
                "properties": {
                    "title": "Recent Errors",
                    "query": "fields @timestamp, @message | filter @message like /ERROR/ | sort @timestamp desc | limit 20",
                    "region": "ap-south-1",
                    "stacked": false,
                    "view": "table"
                }
            }
        ]
    }'

Widget Types

Widget	Purpose	Example
Metric	Line/bar chart of CloudWatch metrics	CPU, memory, request count
Log	Logs Insights query results	Recent errors, slow queries
Alarm	Status of alarms (OK/ALARM)	System health overview
Text	Markdown text	Runbook links, team info
Number	Single metric value (big number)	Current error count

Dashboard Design for On-Call

A production dashboard should answer these questions at a glance:

  ┌─────────────────────────────────────────────────────┐
  │  PRODUCTION OVERVIEW                                 │
  ├─────────────────────┬───────────────────────────────┤
  │  Alarm Status       │  Request Rate                  │
  │  🟢 CPU             │  ████████████  1,200 req/s     │
  │  🟢 Memory          │                                │
  │  🔴 5XX Errors      │  Error Rate                    │
  │  🟢 Latency         │  ██  2.1% (target: <1%)       │
  ├─────────────────────┴───────────────────────────────┤
  │  Latency (p99)                                       │
  │  ▂▃▂▃▂▅▇█▅▃▂▃  ← spike at 10:15                   │
  ├─────────────────────────────────────────────────────┤
  │  ECS Tasks: 5 running / 5 desired                   │
  │  Lambda Errors: 3 in last hour                      │
  ├─────────────────────────────────────────────────────┤
  │  Recent Errors (Log Insights)                       │
  │  10:15:32  ERROR  Connection timeout to RDS         │
  │  10:15:33  ERROR  Connection timeout to RDS         │
  │  10:15:34  ERROR  Connection timeout to RDS         │
  └─────────────────────────────────────────────────────┘

Recommended Dashboards

Dashboard	Audience	Key Widgets
Production Overview	On-call engineer	Alarm status, error rate, latency, recent errors
Cost Dashboard	Engineering manager	Daily spend, service breakdown, forecast
Deployment Dashboard	Release team	Deploy status, rollback count, build times
Infrastructure	Platform team	EC2/ECS utilization, network, storage

Common Mistakes

Mistake	Why It Hurts	Fix
Too many metrics on one chart	Hard to read	Max 4-5 metrics per widget
No alarm status widget	Can’t see overall health at a glance	Add an alarm summary widget at the top
Dashboard per engineer	Inconsistent monitoring	One shared dashboard per environment
No text widget with runbook links	On-call doesn’t know what to do	Add runbook links and escalation info

Quick Reference

Task	Command
Create/update dashboard	aws cloudwatch put-dashboard --dashboard-name NAME --dashboard-body '...'
Get dashboard	aws cloudwatch get-dashboard --dashboard-name NAME
List dashboards	aws cloudwatch list-dashboards
Delete dashboard	aws cloudwatch delete-dashboards --dashboard-names NAME

---

Next → CloudTrail