Basic monitoring - metrics 5 minute interval Detailed Monitoring - metrics 1 minute interval Metrics - CPU, Network, Disk and Status check metrics

Custom metrics

Basic - One minute High resolution - one second EC2 instance should have an IAM role with the permission to push thr metrics

EC2 Included Metrics

  • CPU: CPU utilisation + Credit Usage / Balance
  • Network: In/out
  • Status Check: Health or not VM (1/3) Underlying hardware check (2/3) EBS check (3/3)

Detailed Monitoring

  • To enable detailed monitoring.
  • Go to ec2 monitoring Manage detailed monitoring Enable detailed monitoring

Unified Cloudwatch agent

  • For virtual servers and on prem servers
  • Get additional system logs and metrics
  • The agent can be configured via SSM parameter store
  • Correct IAM permissions should be there

Configuring Agent

  • Install agent
  • Run the wizard
    • Choose OS
    • Ec2 or On prem
    • StatD demon
    • User space
    • interval
  • The wizard will generate a json, which can be stored to ssm parameter store given the necessary permission, and can be loaded back into a fresh VM with fetch-config options via a file or ssm parameter store

Procstat plugin

  • Collect metrics for individual processes (process + stats)
  • Select which process to monitor, via pid, name, pattern
  • Metrics will have prefix prostat
  • CollectD
  • Monitor Host metrics
  • Core level metrics
  • Ec2 dimension
  • High res metrics
  • and a lot
  • Generates a json