EBS loss by tag

Introduction

EBS (Elastic Block Store) loss by tag disrupts the state of EBS volume by detaching it from the node (or EC2) instance using volume ID for a certain duration. In case of EBS persistent volumes, the volumes can self-attach, and the re-attachment step can be skipped.

Use cases

EBS loss by tag tests the deployment sanity (replica availability and uninterrupted service) and recovery workflows of the application pod.

note

• Kubernetes version 1.17 or later is required to execute this fault.
• EBS volume is attached to the instance.
• Appropriate AWS access to attach or detach an EBS volume for the instance.
• The Kubernetes secret should have AWS access configuration (key) in the CHAOS_NAMESPACE. A sample secret file looks like:

  apiVersion: v1
  kind: Secret
  metadata:
    name: cloud-secret
  type: Opaque
  stringData:
    cloud_config.yml: |-
      # Add the cloud AWS credentials respectively
      [default]
      aws_access_key_id = XXXXXXXXXXXXXXXXXXX
      aws_secret_access_key = XXXXXXXXXXXXXXX
• We recommend that you use the same secret name, that is, cloud-secret. Otherwise, you will need to update the AWS_SHARED_CREDENTIALS_FILE environment variable in the fault template and you won't be able to use the default health check probes.
• Go to AWS named profile for chaos to use a different profile for AWS faults and superset permission or policy to execute all AWS faults.
• Go to the common tunables and AWS-specific tunables to tune the common tunables for all faults and AWS-specific tunables.

Below is an example AWS policy to execute the fault.

{
    "Version": "2012-10-17",
    "Statement": [
        {
            "Effect": "Allow",
            "Action": [
                "ec2:AttachVolume",
                "ec2:DetachVolume"
            ],
            "Resource": "*"
        },
        {
            "Effect": "Allow",
            "Action": "ec2:DescribeVolumes",
            "Resource": "*"
        },
        {
            "Effect": "Allow",
            "Action": [
                "ec2:DescribeInstanceStatus",
                "ec2:DescribeInstances"
            ],
            "Resource": "*"
        }
    ]
}

Fault tunables

Mandatory tunables

Tunable	Description	Notes
EBS_VOLUME_TAG	Common tag for target volumes. It'll be in form of key:value (Ex: 'team:devops')
REGION	Region name for the target volumes	For example, us-east-1.

Optional tunables

Tunable	Description	Notes
VOLUME_AFFECTED_PERC	Percentage of total EBS volumes to target	Default: 0 (corresponds to 1 volume), provide numeric value only
TOTAL_CHAOS_DURATION	Time duration for chaos insertion (sec)	Default: 30 s
CHAOS_INTERVAL	The time duration between the attachment and detachment of the volumes (sec)	Default: 30 s
SEQUENCE	Sequence of chaos execution for multiple volumes	Default value: parallel. Supported: serial, parallel
RAMP_TIME	Period to wait before and after injection of chaos in sec	For example, 30

Target single volume

Random EBS volume that is detached from the node, based on the given EBS_VOLUME_TAG tag and REGION region.

The following YAML snippet illustrates the use of this environment variable:

# contains the tags for the EBS volumes 
apiVersion: litmuschaos.io/v1alpha1
kind: ChaosEngine
metadata:
  name: engine-nginx
spec:
  engineState: "active"
  chaosServiceAccount: litmus-admin
  experiments:
  - name: ebs-loss-by-tag
    spec:
      components:
        env:
        # tag of the EBS volume
        - name: EBS_VOLUME_TAG
          value: 'key:value'
        # region for the EBS volume
        - name: REGION
          value: 'us-east-1'
        - name: TOTAL_CHAOS_DURATION
          VALUE: '60'

Target percent of volumes

Percentage of EBS volumes that are detached from the node, based on EBS_VOLUME_TAG tag and REGION region. Tune it by using the VOLUME_AFFECTED_PERC environment variable.

The following YAML snippet illustrates the use of this environment variable:

# target percentage of the EBS volumes with the provided tag
apiVersion: litmuschaos.io/v1alpha1
kind: ChaosEngine
metadata:
  name: engine-nginx
spec:
  engineState: "active"
  chaosServiceAccount: litmus-admin
  experiments:
  - name: ebs-loss-by-tag
    spec:
      components:
        env:
        # percentage of EBS volumes filter by tag
        - name: VOLUME_AFFECTED_PERC
          value: '100'
        # tag of the EBS volume
        - name: EBS_VOLUME_TAG
          value: 'key:value'
        # region for the EBS volume
        - name: REGION
          value: 'us-east-1'
        - name: TOTAL_CHAOS_DURATION
          VALUE: '60'