EC2 network loss

Last updated on Jun 3, 2026

EC2 network loss is an AWS chaos fault that drops a configurable percentage of outbound packets on a target EC2 instance for a configurable duration. The loss can be scoped to specific destination hosts or IPs, and applied on a specific network interface. The fault dispatches the shaping rules via AWS Systems Manager Run Command.

Use this fault to test how a workload reacts when network reliability degrades: do timeouts trigger correctly, do TCP retransmissions absorb the loss transparently, does the application fall back to a healthy replica, do connection pools recover after a churn event?

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:

Partial-loss tolerance: When 10-20% of packets to a critical destination are dropped, does the application stay within SLO?
Full-loss behaviour: When 100% of packets are dropped, does the application treat the destination as down and fall back?
TCP retransmission cost: How much does throughput drop when packet loss is added, and does the application's connection pool absorb it?
Replica failover: When loss to one replica is high, does the client steer traffic to a healthy replica?
Observability: Does the loss show up in network metrics (packets dropped, TCP retransmits) and alerts at the right cadence?

Prerequisites

Kubernetes version: 1.21 or later for the chaos infrastructure cluster.
Target instance is reachable via SSM: The instance has the SSM Agent running and an instance profile with AmazonSSMManagedInstanceCore.
Selector provided: Either EC2_INSTANCE_ID or EC2_INSTANCE_TAG is set.
Network shaping tools installable: tc / iproute2 must be installable on the target. Set INSTALL_NETEM=true (default) to install them if missing.
AWS credentials available: Either an AWS credentials file uploaded as a File Secret in Harness Secret Manager (see Authentication below) or IRSA on the chaos infrastructure service account.

Supported environments

Platform	Support status
Amazon EC2 (Linux instances with SSM Agent)	Supported
Amazon EKS managed worker nodes	Supported (if SSM Agent is installed)
Amazon EKS self-managed worker nodes	Supported (if SSM Agent is installed)
Targeting by tag	Supported via `EC2_INSTANCE_TAG`
Targeting by ID	Supported via `EC2_INSTANCE_ID`
Windows instances	Not supported (Linux-only network shaping)

Permissions required

{
  "Version": "2012-10-17",
  "Statement": [
    {
      "Effect": "Allow",
      "Action": [
        "ec2:DescribeInstances",
        "ec2:DescribeInstanceStatus"
      ],
      "Resource": "*"
    },
    {
      "Effect": "Allow",
      "Action": [
        "ssm:SendCommand",
        "ssm:CancelCommand",
        "ssm:GetCommandInvocation",
        "ssm:DescribeInstanceInformation",
        "ssm:GetDocument",
        "ssm:DescribeDocument"
      ],
      "Resource": "*"
    }
  ]
}

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

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 File Secret should contain an 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.

Fault tunables

Required parameters

Tunable	Description	Default
`REGION`	AWS region that hosts the target instance.	(required)
`EC2_INSTANCE_ID` or `EC2_INSTANCE_TAG`	One of these must be set to select the target instance(s).	`""`

Chaos parameters

Tunable	Description	Default
`NETWORK_PACKET_LOSS_PERCENTAGE`	Percentage of outbound packets to drop (0-100).	`100`
`NETWORK_INTERFACE`	Network interface to apply shaping on.	`eth0`
`DESTINATION_HOSTS`	Comma-separated list of destination hostnames to target. Leave empty to affect all destinations.	`""`
`DESTINATION_IPS`	Comma-separated list of destination IPs/CIDRs to target. Leave empty to affect all destinations.	`""`
`TOTAL_CHAOS_DURATION`	Duration of the fault in seconds.	`30`
`INSTANCE_AFFECTED_PERC`	Percentage of matching instances to target (only with `EC2_INSTANCE_TAG`). `0` targets one instance.	`0`
`INSTALL_DEPENDENCIES`	Install dependencies inside the target instance if missing. Set to `False` to skip.	`True`
`INSTALL_NETEM`	Install the `tc` / netem networking tools if missing.	`true`
`PROXY`	HTTP/HTTPS proxy used by the in-instance installer (for example `https_proxy=http://proxy.server:3128`).	`""`
`SEQUENCE`	Order in which multiple instances are processed: `parallel` or `serial`.	`parallel`
`RAMP_TIME`	Wait period in seconds before and after the fault.	`0`

Authentication

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

Scope the blast radius with destinations

Without DESTINATION_HOSTS or DESTINATION_IPS, packet loss applies to all outbound traffic, including the SSM control channel back to AWS. Scope to specific destinations whenever possible.

Fault execution in brief

Sends an SSM Run Command to the selected instance(s) in REGION that adds a network-shaping rule on NETWORK_INTERFACE to drop NETWORK_PACKET_LOSS_PERCENTAGE percent of outbound packets to DESTINATION_HOSTS/DESTINATION_IPS for TOTAL_CHAOS_DURATION seconds, then removes the rule.

Expected behavior during fault execution

Outbound packets to the targeted destinations are dropped at the specified percentage.
TCP connections absorb small loss percentages (1-5%) through retransmits, with reduced throughput.
Larger loss percentages (≥20%) cause noticeable slowdown and connection drops.
100% loss makes destinations appear unreachable; existing connections eventually time out and new ones fail to establish.
Network counters on the target instance (netstat -s, kernel TCP retransmits) reflect the loss.

When the fault ends

The chaos pod removes the network-shaping rule. Traffic returns to baseline immediately, but connections that failed during the fault must be re-established by the caller.

Signals to watch

End-to-end availability: Use an HTTP probe against an endpoint served by the affected destination.
Synthetic packet loss: Use a command probe that runs ping -c 20 <destination> via SSM during the fault and asserts loss is close to NETWORK_PACKET_LOSS_PERCENTAGE.
TCP retransmissions: Use a Prometheus probe on node_netstat_Tcp_RetransSegs to confirm the loss surfaces as retransmits.

Verify the fault execution effect

While the experiment is running:

Confirm the shaping rule is in place via SSM.

aws ssm send-command \
  --region <region> \
  --instance-ids <id> \
  --document-name AWS-RunShellScript \
  --parameters 'commands=["tc qdisc show dev <NETWORK_INTERFACE>"]'

Measure loss to the destination.

aws ssm send-command \
  --region <region> \
  --instance-ids <id> \
  --document-name AWS-RunShellScript \
  --parameters 'commands=["ping -c 20 <destination-host>"]'

Recovery and cleanup

End of duration: The chaos pod removes the network-shaping rule.
Abort the experiment: Stopping the experiment from Chaos Studio cancels the in-flight SSM command and removes the rule.
Manual cleanup: If a rule is left behind, remove it via SSM: tc qdisc del dev <NETWORK_INTERFACE> root.

Limitations

Linux-only payload: This fault shapes traffic on Linux instances. For Windows hosts, use windows-ec2-network-loss.
SSM Agent required: Instances without the SSM Agent online cannot be targeted.
Outbound only: Loss is applied to traffic leaving NETWORK_INTERFACE. Inbound loss is not directly controllable from the receiver side.
Control-channel impact: Without DESTINATION_HOSTS/DESTINATION_IPS, the rule affects all outbound traffic including the SSM control channel; very high loss values may delay cleanup.
DNS resolution timing: Hosts in DESTINATION_HOSTS are resolved once at the start of the fault.

Troubleshooting

EC2 network loss experiment fails with InvalidInstanceId in Harness Chaos Engineering

The SSM Agent is not online for the target instance. Confirm with aws ssm describe-instance-information --filters 'Key=InstanceIds,Values=<id>'. If missing, install the SSM Agent and attach an instance profile that includes AmazonSSMManagedInstanceCore.

EC2 network loss runs but no packet loss is observed

The most common causes are: tc/iproute2 failed to install (set INSTALL_NETEM=true and verify network egress, or use PROXY); NETWORK_INTERFACE is wrong (use 'ip link show' to confirm); or DESTINATION_HOSTS resolved to IPs that the target reaches via a different interface. Run 'tc qdisc show dev <interface>' via SSM during the fault and use 'ping' to measure loss directly.

EC2 network loss disrupted my SSM connection and the fault hangs

DESTINATION_HOSTS / DESTINATION_IPS were left empty, so loss applied to all outbound traffic including the SSM control channel. Wait for TOTAL_CHAOS_DURATION to elapse and the cleanup command to land, then rerun with explicit DESTINATION_HOSTS or DESTINATION_IPS that exclude the SSM endpoint.

EC2 network latency: Add latency instead of dropping packets.
EC2 DNS chaos: Block DNS resolution for specific hostnames.
Windows EC2 network loss: Packet loss for Windows instances.
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​