ECS container HTTP reset peer

Last updated on Jun 5, 2026

ECS container HTTP reset peer is an AWS chaos fault that resets HTTP connections to clients on TARGET_SERVICE_PORT after RESET_TIMEOUT milliseconds inside a percentage of running ECS tasks (EC2 launch type) for a configurable duration. The fault interposes a transparent HTTP proxy on the container instance, scoped to the target container via ECS container metadata and dispatched via AWS Systems Manager Run Command. After RESET_TIMEOUT ms the proxy sends a TCP RST to the client.

Use this fault to test how HTTP clients behave when the server abruptly resets the connection mid-request: whether retries engage, whether the client surfaces a clear error, and whether downstream systems are protected from cascading failures.

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:

Reset handling: When the server resets the connection mid-request, does the client surface a clear error to its caller, or does it retry forever?
Idempotency safety: For non-idempotent requests that the client retries on reset, does the server detect and reject duplicate requests?
Connection pool churn: Do reset connections cause the client's connection pool to churn, leading to TLS handshake storms?
Long-lived streams: Do gRPC streams and WebSocket connections recover, or do they degrade gracefully?
Monitoring fidelity: Do client-side EOF and connection reset by peer metrics rise as expected, and do alarms fire?

Prerequisites

Kubernetes version: 1.21 or later for the chaos infrastructure cluster. Go to What's supported to confirm distribution support.
Target ECS service or cluster: CLUSTER_NAME exists in REGION and uses the EC2 launch type.
Container instances are SSM-managed.
ECS container metadata enabled.
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	Supported
Amazon ECS on Fargate launch type	Not supported
HTTP/1.1 and HTTP/2 traffic on `TARGET_SERVICE_PORT`	Supported
HTTPS terminated inside the container	Not supported
Linux container instances	Supported
Windows container instances	Not supported
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:ListTasks",
        "ecs:ListContainerInstances",
        "ecs:DescribeContainerInstances"
      ],
      "Resource": "*"
    },
    {
      "Effect": "Allow",
      "Action": [
        "ssm:SendCommand",
        "ssm:CancelCommand",
        "ssm:GetCommandInvocation",
        "ssm:DescribeInstanceInformation"
      ],
      "Resource": "*"
    },
    {
      "Effect": "Allow",
      "Action": [
        "ec2:DescribeInstances",
        "ec2messages:AcknowledgeMessage",
        "ec2messages:DeleteMessage",
        "ec2messages:FailMessage",
        "ec2messages:GetEndpoint",
        "ec2messages:GetMessages",
        "ec2messages:SendReply"
      ],
      "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 HTTP reset peer 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.	`""`
`TASK_REPLICA_ID`	ID of a specific task replica to target.	`""`
`TASK_REPLICA_AFFECTED_PERC`	Percentage of running tasks to affect when `SERVICE_NAME` is set.	`100`

Chaos parameters

Tunable	Description	Default
`RESET_TIMEOUT`	Number of milliseconds the proxy waits after the client connects before sending a TCP RST. Use `0` to reset immediately on connect.	`10000`
`TARGET_SERVICE_PORT`	TCP port the affected container serves on.	`80`
`PROXY_PORT`	Port the chaos proxy listens on inside the container.	`20000`
`NETWORK_INTERFACE`	Network interface inside the container on which to install the redirect rules. `auto` discovers the primary interface.	`auto`
`TOTAL_CHAOS_DURATION`	Duration of the fault in seconds.	`60`
`INSTALL_DEPENDENCIES`	Install the proxy and networking tooling on each container instance if missing.	`true`
`SEQUENCE`	Order in which multiple tasks are affected: `parallel` installs the proxy on all selected tasks at once; `serial` does so one at a time.	`parallel`
`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 (or the explicit TASK_REPLICA_ID), picks TASK_REPLICA_AFFECTED_PERC of them, and dispatches a transparent HTTP proxy via AWS Systems Manager Run Command into the affected container. New TCP connections on TARGET_SERVICE_PORT are redirected through the proxy on PROXY_PORT; the proxy waits RESET_TIMEOUT ms after the connection is established and then sends a TCP RST to the client. After TOTAL_CHAOS_DURATION seconds the proxy is stopped and the redirect rules are removed.

Expected behavior during fault execution

HTTP clients connecting to the affected containers experience connection reset by peer errors after RESET_TIMEOUT ms.
Idempotent clients with retry will retry; non-idempotent clients surface the error to the caller.
TCP connections established before the chaos started are not affected (security-group-style; the proxy only intercepts new connections).
CloudWatch metrics may show elevated 5xx, 502, or 0-length responses (depending on which side of the reset the load balancer is).

When the fault ends

The chaos pod stops the proxy and removes the redirect rules. New connections are not reset.

Signals to watch

Attach resilience probes to assert each layer:

Connection error counters: Use a Prometheus probe on client-side connection-reset counters.
Application availability: Use an HTTP probe and confirm the failure surfaces.
Retry counters: Use a Prometheus probe on retry counters to detect amplification.
Load balancer 5xx: Use a Prometheus probe on aws_applicationelb_httpcode_target_5xx_count to confirm the LB sees the failures.

Verify the fault execution effect

While the experiment is running, confirm connections are reset:

Probe with curl.
```
curl --max-time 30 -v http://<service-endpoint>:<port>/
```
You should see Empty reply from server or Connection reset by peer after approximately RESET_TIMEOUT ms.

Inspect the redirect rules on the host (via SSM).

aws ssm send-command \
  --region <region> \
  --document-name AWS-RunShellScript \
  --instance-ids <container-instance-id> \
  --parameters 'commands=["iptables -t nat -L PREROUTING -n"]'

Inspect SSM command status.

aws ssm list-command-invocations --region <region> --details --filters "key=Status,value=InProgress"

Recovery and cleanup

End of duration: The chaos pod stops the proxy and removes the redirect rules on each host.
Abort the experiment: Stopping the experiment from Chaos Studio cancels the SSM command and removes the proxy.
Manual recovery: If the fault exits before cleanup, remove the redirect rules and kill the proxy process via SSM.
Workload recovery: Connections reset during the chaos window are not restored; clients must re-establish them.

Limitations

EC2 launch type only.
Container metadata must be enabled.
SSM-managed hosts only.
Linux-only.
HTTPS terminated inside the container is not supported.
Only new connections are affected: Already-established connections continue normally.
One port per experiment: The proxy intercepts only TARGET_SERVICE_PORT.
Cross-region targeting: A single experiment targets one region (the value of REGION).

Troubleshooting

ECS container HTTP reset peer fails with AccessDeniedException in Harness Chaos Engineering

The credentials supplied to the chaos pod do not have the required ECS or SSM permissions. Confirm the IAM policy attached to the user, role, or IRSA service account includes ecs:DescribeServices, ecs:DescribeTasks, ssm:SendCommand, and ssm:GetCommandInvocation.

Reset is not observed by client

The most common causes are: the client connects on a different port (set TARGET_SERVICE_PORT correctly); the client used a pre-existing pooled connection that escaped the proxy; the load balancer routed traffic to a non-affected task; or the proxy could not install the redirect rule (check NETWORK_INTERFACE). Force a fresh connection from the client and re-check.

Tasks are restarted by health checks during the chaos window

If the health check uses TARGET_SERVICE_PORT, the proxy resets it too, causing the probe to fail. Either point the health check at a separate port not intercepted by the proxy, or increase the health check failure threshold during the experiment so probes can tolerate the resets.

Redirect rules remain after the chaos window

If the cleanup SSM command failed, the iptables PREROUTING redirect rules may persist on the container instance. Send an SSM AWS-RunShellScript command to the affected container instance that flushes the iptables NAT PREROUTING chain. The exact command is recorded in the chaos pod logs.

ECS container HTTP latency: Delay the response instead of resetting.
ECS container HTTP status code: Return a chaos status code instead of resetting.
ECS container HTTP modify body: Modify the body instead of resetting.
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​