Skip to main content
Understanding the Amazon EC2 instance lifecycle is crucial for managing compute resources efficiently. This guide walks you through each state an EC2 instance passes—from launch to termination—highlighting resource allocation, billing implications, and transition triggers.

​
EC2 Instance States

Stopping an instance does not delete attached EBS volumes or Elastic IP addresses—you will continue to incur charges for those resources until you release them.
The image illustrates the EC2 Instance Life Cycle, showing the stages from launch to termination, including states like pending, running, stopping, and stopped. It also highlights the differences in resources between running and stopped instances.

​
Detailed State Transitions

  1. Pending
    • AWS locates capacity for your chosen instance type.
    • Resources are reserved, networking is configured, and the hypervisor initializes the VM.
  2. Running
    • Health checks complete and instance-store volumes attach.
    • A public IP is assigned (if enabled at subnet level).
    • Your applications can now serve traffic.
  3. Stopping
    • Issued via Console, AWS CLI (aws ec2 stop-instances), or within the guest OS.
    • The guest OS performs a clean shutdown; the hypervisor marks resources for release.
  4. Stopped
    • Compute resources are released, halting billing for CPU and memory.
    • Storage (EBS), private IP, and Elastic IP remain allocated.
  5. Shutting-down
    • Occurs when you terminate an instance from either the running or stopped state.
    • AWS deallocates virtual hardware; cleanup operations begin.
  6. Terminated
    • The instance record is removed from your account.
    • Instance-store volumes and default EBS volumes flagged for deletion are erased.

​
Edge Cases and Best Practices

Sending a terminate command is irreversible. Ensure you’ve backed up any critical data on instance-store volumes or unencrypted EBS volumes before terminating.

​
Additional Resources

Watch Video