ECS container volume detach

Last updated on Jun 5, 2026

ECS container volume detach is an AWS chaos fault that detaches the data EBS volume attached to a percentage of running ECS tasks (filtered by compatibility through REQUIRED_COMPATIBILITIES) for a configurable duration, then reattaches the volume when the fault ends. While the volume is detached the application loses access to its mounted data; reads and writes against the detached path fail.

Use this fault to test how a workload behaves when its mounted storage briefly disappears: whether the application surfaces a clear error, whether stateful services (databases, queues) recover cleanly when the volume returns, and whether your monitoring detects the storage outage quickly.

Run your first experiment

If you have not configured the chaos infrastructure yet, go to Quickstart to install the chaos infrastructure and run an experiment end to end.

Use cases

Run this fault when you want to answer concrete questions like:

Stateful workload resilience: When the data volume disappears, does the application fail fast with a clear error or hang waiting on IO?
Recovery after reattach: When the volume returns, does the application reattach to the mount cleanly, or does it require a restart?
Replication failover: For stateful services with replicas, does a healthy replica take over when its peer loses storage?
Storage monitoring: Does CloudWatch (or your monitoring) detect the volume outage within the alarm SLA?
Backup readiness: Is the runbook to restore from snapshot accurate and timed correctly when a volume becomes unreachable in production?

Prerequisites

Kubernetes version: 1.21 or later for the chaos infrastructure cluster. Go to What's supported to confirm distribution support.
Target ECS service: CLUSTER_NAME exists in REGION. If SERVICE_NAME is set, the fault selects from that service's running tasks.
Tasks use an attachable volume: The task definition mounts an EBS volume (Fargate task volume support) or the EC2 launch type uses an EBS-backed bind mount that can be detached at the EC2 level.
AWS credentials available: Either an AWS credentials file uploaded as a File Secret in Harness Secret Manager (see Authentication below) or an IAM role for service accounts (IRSA) bound to the chaos infrastructure service account.
IAM permissions granted: The credentials or role include the permissions listed below.

Supported environments

Platform	Support status
Amazon ECS on EC2 launch type with EBS-backed volumes	Supported
Amazon ECS on Fargate launch type with EBS volume attachments	Supported
Tasks without attached EBS volumes	Not supported (nothing to detach)
AWS regions	Supported in every commercial region; pass the region in `REGION`

Permissions required

The IAM principal that the chaos pod uses (the credentials mounted from the Harness Secret Manager file secret, the IRSA role on the chaos service account, or the role assumed via ASSUME_ROLE_ARN) needs the following AWS actions.

{
  "Version": "2012-10-17",
  "Statement": [
    {
      "Effect": "Allow",
      "Action": [
        "ecs:DescribeClusters",
        "ecs:DescribeServices",
        "ecs:DescribeTasks",
        "ecs:DescribeTaskDefinition",
        "ecs:ListTasks"
      ],
      "Resource": "*"
    },
    {
      "Effect": "Allow",
      "Action": [
        "ec2:DescribeVolumes",
        "ec2:DescribeInstances",
        "ec2:AttachVolume",
        "ec2:DetachVolume"
      ],
      "Resource": "*"
    }
  ]
}

Go to common policy for all AWS faults to use a single superset IAM policy across every AWS fault.

Authentication

The fault supports three credential delivery models. Pick one based on how your chaos infrastructure is deployed.

Method	When to use it	How to configure
Harness Secret Manager file secret	Chaos infrastructure runs outside EKS, or you want explicit static credentials	Upload the AWS credentials file as a File Secret in Harness Secret Manager and reference its identifier via `AWS_AUTHENTICATION_SECRET`
IAM Roles for Service Accounts (IRSA)	Chaos infrastructure runs in EKS and uses an OIDC-bound service account	No tunable changes; the chaos pod inherits the role automatically. Go to AWS IAM integration to set it up
Assume role	The fault needs to act in a different account or with elevated permissions	Set `ASSUME_ROLE_ARN` to the role ARN; the chaos pod assumes the role on top of its base credentials

When using the Harness Secret Manager method, the contents of the File Secret should be the AWS credentials file in the standard ~/.aws/credentials format:

[default]
aws_access_key_id = REPLACE_WITH_ACCESS_KEY_ID
aws_secret_access_key = REPLACE_WITH_SECRET_ACCESS_KEY

Upload this file as a File Secret in Harness Secret Manager (Project Setup → Secrets → New File Secret), and pass the secret identifier in AWS_AUTHENTICATION_SECRET when configuring the fault.

Go to AWS named profile for chaos to switch between profiles inside a single credentials file.

Fault tunables

Configure the following fault parameters when you add ECS container volume detach to an experiment in Chaos Studio. Defaults are shown for reference.

Required parameters

Tunable	Description	Default
`CLUSTER_NAME`	Name of the target ECS cluster.	(required)
`REGION`	AWS region that hosts the ECS cluster (for example `us-east-1`).	(required)

Targeting parameters

Tunable	Description	Default
`SERVICE_NAME`	Name of the target ECS service. When set, the fault selects `TASK_REPLICA_AFFECTED_PERC` of the service's running tasks.	`""`
`REQUIRED_COMPATIBILITIES`	Comma-separated list of compatibilities (`fargate`, `ec2`) that a task must support to be considered.	`fargate,ec2`
`TASK_REPLICA_AFFECTED_PERC`	Percentage of qualifying tasks to detach volumes from.	`100`

Chaos parameters

Tunable	Description	Default
`TOTAL_CHAOS_DURATION`	Duration of the fault in seconds. The volume stays detached for this period.	`30`
`CHAOS_INTERVAL`	Delay in seconds between successive iterations when running for more than one cycle.	`30`
`RAMP_TIME`	Wait period in seconds before and after the fault. Go to ramp time to read how it is applied.	`0`

Authentication

Tunable	Description	Default
`ASSUME_ROLE_ARN`	ARN of an IAM role to assume on top of the base credentials. Leave empty to use the base credentials directly.	`""`
`AWS_AUTHENTICATION_SECRET`	Identifier of the File Secret in Harness Secret Manager that contains the AWS credentials file. Not required when using IRSA.	`""`

Tunables that apply to every fault are documented in common tunables for all faults. AWS-specific shared tunables are documented in common AWS fault tunables.

Fault execution in brief

Resolves the running tasks for SERVICE_NAME filtered by REQUIRED_COMPATIBILITIES, picks TASK_REPLICA_AFFECTED_PERC of them, and calls DetachVolume on the EBS volume attached to each selected task. After TOTAL_CHAOS_DURATION seconds, the chaos pod calls AttachVolume to restore each volume to its original device.

Expected behavior during fault execution

The target volumes transition from in-use to detaching to available.
Reads and writes against the mount point inside the affected containers return IO errors.
Stateful applications that depend on the volume (databases, queues) may exit, restart, or enter a degraded mode.
ECS may attempt to restart the task if the application exits; without the volume, restarts may fail.
CloudWatch EBS volume metrics show the detach.

When the fault ends

The chaos pod calls AttachVolume to reattach each volume to its original device. The application may need to remount or restart to see the returned volume.

Signals to watch

Attach resilience probes to assert each layer:

Volume state: Use a command probe that runs aws ec2 describe-volumes --volume-ids <id> and asserts on State.
Application errors: Use a Prometheus probe on application-level IO error counters.
Task lifecycle: Use a command probe that runs aws ecs describe-tasks and asserts on lastStatus and desiredStatus.
Replication health: For replicated stateful workloads, use a Prometheus probe on replication-lag metrics.

Verify the fault execution effect

While the experiment is running, confirm the volume is detached and then reattached:

Inspect volume state.

aws ec2 describe-volumes \
  --region <region> \
  --volume-ids <vol-id> \
  --query "Volumes[].[VolumeId,State,Attachments[].Device]"

During the fault State should be available (detached). After recovery it should be in-use again.

Inspect task state.

aws ecs describe-tasks --cluster <cluster> --tasks <arn> --region <region> \
  --query "tasks[].[lastStatus,desiredStatus]"

The task may be restarted; the resulting state depends on the application.

Check application logs.

ECS task logs in CloudWatch should show IO errors during the chaos window and recovery messages after the volume returns.

Recovery and cleanup

End of duration: The chaos pod calls AttachVolume for every detached volume.
Abort the experiment: Stopping the experiment from Chaos Studio also triggers the reattach.
Manual recovery: If the fault exits before reattach, run aws ec2 attach-volume --volume-id <vol-id> --instance-id <i> --device <device> with the values recorded in the chaos pod logs.
Workload recovery: Depending on the application, you may need to remount the filesystem or restart the task.

Limitations

Volume must be attachable: Only EBS volumes that support detach/attach are affected; instance-store and ephemeral storage are not detached by this fault.
Multi-attach behaviour: If the volume uses Multi-Attach, detaching from one instance leaves the volume available on the other; the fault still records the original attachment for reattach.
AZ constraints on reattach: EBS volumes can only attach to instances in the same AZ; this fault preserves the original instance, so AZ constraints are honoured automatically.
Cross-region targeting: A single experiment targets one region (the value of REGION).

Troubleshooting

ECS container volume detach fails with VolumeInUse or InvalidVolume.NotFound

VolumeInUse means another operation is in progress on the volume (a snapshot, another detach, or an attachment change). Wait a few seconds and retry. InvalidVolume.NotFound usually means the volume was deleted between discovery and the detach call; the most common cause is a parallel autoscaling or replacement event. Inspect 'aws ec2 describe-volumes' and rerun.

Volume reattach fails after the chaos window

The most common causes are: the original instance was replaced during the chaos window (autoscaling, manual termination), so the device path is no longer valid; or another volume was attached to the same device path while the chaos volume was detached. Reattach the volume to the new instance manually with 'aws ec2 attach-volume' to an unused device path.

Application does not see the volume after reattach

Even after reattach, the OS may not automatically remount the filesystem. SSH or SSM into the container instance and run 'mount -a' (Linux) or restart the affected task so the mount is re-established. For stateful applications, follow the application's documented restart procedure to ensure clean recovery.

ECS container volume detach reports no candidate volumes were found

The most common causes are: SERVICE_NAME does not have tasks that match REQUIRED_COMPATIBILITIES; the task definition does not declare any EBS volume mounts; or the fault could not resolve the underlying volume ID from the task definition. Confirm with 'aws ecs describe-tasks' and 'aws ecs describe-task-definition' that the tasks expose an EBS volume.

EBS loss by ID: Detach a specific EBS volume outside an ECS context.
EBS loss by tag: Detach EBS volumes selected by tag.
ECS container IO stress: Stress filesystem IO instead of detaching the volume.
Common AWS fault tunables: Shared environment variables for AWS faults.

Use cases​

Prerequisites​

Supported environments​

Permissions required​

Authentication​

Fault tunables​

Fault execution in brief​

Expected behavior during fault execution​

Signals to watch​

Verify the fault execution effect​

Recovery and cleanup​

Limitations​

Troubleshooting​

Related faults​