EC2 status check failed: system vs instance vs EBS
A failed EC2 status check is either AWS's problem (system) or yours (instance). Read which check failed, then stop and start, reboot, or fix the OS.
What you see
- Instance state
impairedwithStatusCheckFailed_System,StatusCheckFailed_InstanceorStatusCheckFailed_AttachedEBSraised - The instance is unreachable but shows
running
What it means
EC2 runs status checks every minute. All pass means OK. A failure means impaired. The check that failed tells you who has to act.
A system check failure points at the AWS side: the host's power, network, or software. An instance check failure points at your side: the OS, its network config, memory, or file system.
How to confirm the cause
- 1See which check is failing
aws ec2 describe-instance-status --instance-ids INSTANCE_ID --include-all-instances
Compare
SystemStatusandInstanceStatus. Attached EBS and performance checks are reported separately. - 2Read the console output for boot errors
aws ec2 get-console-output --instance-id INSTANCE_ID --output text
Useful for instance check failures: kernel panics, file system errors, failed mounts.
- 3Check the CloudWatch metrics
# Metrics: StatusCheckFailed_System, StatusCheckFailed_Instance, StatusCheckFailed_AttachedEBS
Create alarms on these so you find out before users do.
Causes and fixes
System status check: the underlying host
- HOW TO TELL
StatusCheckFailed_Systemis raised.- ROOT CAUSE
- Loss of network connectivity, loss of system power, or software or hardware issues on the physical host, all of which need AWS to repair.
- FIX
- For an EBS-backed instance, stop and start it. In most cases this moves it to a new host. For an instance-store instance, terminate and replace it, because instance store data is lost when it stops.
aws ec2 stop-instances --instance-ids INSTANCE_ID aws ec2 start-instances --instance-ids INSTANCE_ID
Instance status check: your OS or config
- HOW TO TELL
StatusCheckFailed_Instanceis raised.- ROOT CAUSE
- The checks send an ARP request to the network interface. Failure points at incorrect networking or startup configuration, exhausted memory, a corrupted file system, or an incompatible kernel.
- FIX
- Reboot first. If it persists, read the console output and fix the cause, for example repair the file system by attaching the volume to another instance.
Attached EBS status check
- HOW TO TELL
StatusCheckFailed_AttachedEBSis raised (Nitro instances only).- ROOT CAUSE
- One or more attached EBS volumes cannot complete I/O: a problem in the storage layer, the host, or the connection between them.
- FIX
- Wait for AWS to resolve it, replace the volume, or stop and start the instance to move it to a new host if the workload is not driving I/O.
How to stop it coming back
- Alarm on
StatusCheckFailedand enable automatic instance recovery where it suits. - Use Auto Scaling groups so a bad instance is replaced for you.
- Keep data off instance store volumes unless it is disposable.
Keep going
- Runbook: EC2 SSH: Connection timed out
- Runbook: No space left on device
- Cheat sheet: AWS CLI
- Lesson: the cloud topic in Learn
Sources
AWS documentation (docs.aws.amazon.com), read 18 Nov 2026.