CloudWatch

• System Metrics: This are metrics that you get out of the box. The more managed the service is, the more you get.

• Application Metrics: By installing the CloudWatch agent you get information from inside your EC2 instances

• Alarms: alerts when something goes wrong.

Default metrics are:

• CPU Utilisation

• Network Throughput

Custom metrics (requiring CloudWatch agent) are:

• EC2 Memory Utilization

• ESB Storage Capacity

CloudWatch default metrics

EC2 (Think "CPU-DNS"):

• CPU: CPUUtilization (%)

• Disk: DiskRead/Write (Ops and Bytes)

• Network: NetworkIn/Out

• Status: StatusCheckFailed

EBS (Think "VRW"):

• Volume Read/Write (Bytes & Ops)

• VolumeQueueLength

RDS (Think "CPU-FD"):

• CPU Usage

• Free Memory

• DB Connections

• Free Storage

Load Balancer (Think "RHE"):

• Request Count

• Healthy/Unhealthy Hosts

• Error codes (4XX/5XX)

Lambda (Think "TED"):

• Throttles

• Errors

• Duration

Remember: Default monitoring is 5 min, detailed is 1 min (costs extra).