It provides metrics for every service in AWS. Metric is a variable to monitor like CPU utilization, NetworkIN etc.

  • Metrics belong to namespaces
  • Dimension is an attribute of metric (instance id, environment…)

Standard monitoring is every 5 minutes, detailed monitoring every 1 minute

Metric Stream

Continuously stream CloudWatch metrics to a destination, with near-real-time delivery and low latency.

  • Kinesis Data Firehose (and then its destinations)
  • 3rd party: Datadog, Dynatrace, New Relic, Splunk, Sumo Logic…

Metrics

Exist both in standard and detailed monitoring:

CategoryMetric NameWhat it Tracks
CPUCPUUtilizationPercentage of allocated CPU being used.
CPUCreditUsage / BalanceOnly for burstable instances (T2, T3, etc.).
NetworkNetworkIn / NetworkOutTotal Bytes received/sent on all interfaces.
NetworkPacketsIn / OutTotal number of packets received/sent.
Disk (Instance Store)DiskReadOps / WriteOpsI/O operations on ephemeral (local) disks.
DiskReadBytes / WriteBytesData throughput on ephemeral (local) disks.
Status ChecksStatusCheckFailedReports if the instance or system is failing.
StatusCheckFailed_InstanceFailure due to your OS/Configuration.
StatusCheckFailed_SystemFailure due to AWS hardware/power issues.

Exists only in detailed monitoring:

  • Memory (RAM): mem_used_percent, mem_available, etc.

  • Disk Utilization: disk_used_percent (how full your actual drive/partition is).

  • Swap Usage: swap_used.

  • Process Count: Number of active threads/tasks.