CloudConsole

aws ec2 reboot-instances

Updated 16 November 2026 · How we write and check guides

Short answer

aws ec2 reboot-instances --instance-ids i-0123456789abcdef0 queues an operating-system reboot. The instance stays on the same host and keeps its public IP, instance store data and billing period; if the OS will not shut down cleanly within a few minutes, EC2 forces a hard reboot.

Terminal running aws ec2 reboot-instances --instance-ids…
A worked example from this page. Copy commands from the text below, not from the image.

What the command does

reboot-instances is the API equivalent of running sudo reboot, except it works when you cannot log in. EC2 signals the guest OS to restart; the instance never leaves the running state, so you will not see stopping or pending in describe-instances.

Because it stays on the same physical host, a reboot does not fix hardware or host problems. If the system status check is failing, you need a stop and start so the instance can move.

Syntax

aws ec2 reboot-instances --instance-ids i-0123456789abcdef0 [i-0fedcba9876543210 ...] [--dry-run]

On success the command prints nothing at all. That catches people out: no output and exit code 0 means the reboot request was accepted.

Worked examples

Reboot and confirm the instance came back healthy:

aws ec2 reboot-instances --instance-ids i-0123456789abcdef0
sleep 60
aws ec2 wait instance-status-ok --instance-ids i-0123456789abcdef0 && echo healthy

The pause matters: immediately after the request, status checks may still report the pre-reboot result, so the waiter can return before the reboot has even happened.

See what happened during boot if it does not come back:

aws ec2 get-console-output --instance-id i-0123456789abcdef0 --latest --output text

Reboot a group of instances by tag (one at a time keeps a service available):

for id in $(aws ec2 describe-instances \
    --filters "Name=tag:Role,Values=worker" "Name=instance-state-name,Values=running" \
    --query "Reservations[].Instances[].InstanceId" --output text); do
  aws ec2 reboot-instances --instance-ids "$id"
  sleep 60
  aws ec2 wait instance-status-ok --instance-ids "$id"
done

Confirm who rebooted an instance using CloudTrail:

aws cloudtrail lookup-events \
  --lookup-attributes AttributeKey=EventName,AttributeValue=RebootInstances \
  --query "Events[].[EventTime,Username]" --output table

Reboot vs stop/start

reboot-instancesstop + start
Physical hostSameUsually different
Public IPv4 (non-Elastic)KeptChanges
Instance store dataKeptLost
Fixes failed system status checkNoUsually
BillingContinues uninterruptedStops, then restarts with a new minimum period
Instance type change possibleNoYes, while stopped

Gotchas

  • A hung kernel may ignore the signal. EC2 falls back to a hard reboot after a few minutes, so be patient before escalating to stop/start.
  • Load balancer health checks will fail briefly. Expect targets to go unhealthy during the reboot; deregister first if you need zero errors.
  • A failing instance status check (kernel panic, full disk, bad fstab) often comes straight back after a reboot. Read the console output before retrying.
  • Termination and stop protection do not block reboots. They only guard against terminate and stop.
  • Only API reboots are in CloudTrail. A reboot started from inside the operating system, such as sudo reboot or an unattended-upgrades restart, never calls the EC2 API, so CloudTrail has no record of it. Check the OS journal or the console output for those.

Questions

How do I reboot an EC2 instance from the terminal?

Run aws ec2 reboot-instances --instance-ids i-0123456789abcdef0. If you are logged in to the instance itself, sudo reboot has the same effect.

Does rebooting an EC2 instance change its IP address?

No. A reboot keeps both the private and public IP addresses. Only a stop and start releases an auto-assigned public IP.

Why did reboot-instances print nothing?

That is normal. The API returns an empty response on success, so the CLI prints nothing; check status checks afterwards to confirm the instance recovered.

When this happens away from your desk

A reboot is often the first thing tried when an alarm fires at night. CloudConsole, a phone app for AWS, is built to show the failing status check and the instance's alarms on a phone so you can judge whether a reboot is the right move.

About the app → · Download on the App Store

Sources