Skip to content

EC2 Instances

EC2 Instances in Amazon Web Services can become a significant source of waste, particularly when they are underutilized or completely idle.

We identify this waste through an advanced detection algorithm that analyzes multiple CloudWatch signals over an extended observation period. Rather than relying on a single metric, we correlate several dimensions of instance activity:

  • CPU utilization — both sustained averages and peak activity, ensuring periodic workloads like cron jobs are not falsely flagged
  • Network activity — inbound and outbound traffic patterns to detect instances with no meaningful communication
  • Disk I/O — read and write operations to catch processing-heavy workloads that may have low network footprints

Only instances where all signals consistently indicate inactivity are flagged as unused — minimizing false positives and giving you high-confidence results.

For flagged instances, we also provide right-sizing recommendations based on observed usage patterns, suggesting optimized instance families (including Graviton migration opportunities) and appropriate sizes to help reduce costs even for instances you choose to keep running.

EC2 Dedicated Hosts

EC2 Dedicated Hosts in Amazon Web Services represent a substantial cost risk, particularly when they are idle without any instances running on them.

By actively monitoring and assessing EC2 Dedicated Hosts for signs of non-use, you can make informed decisions to release these hosts and avoid unnecessary expenses.

Adopting this vigilant approach ensures more efficient use of resources and promotes cost savings within your AWS environment.