Amazon ECS Container Agent

Amazon Elastic Container Service Agent
Under Apache License 2.0
By aws

go amazon-ec2 docker-container amazon-ecs-agent amazon-linux-ami

Amazon ECS Container Agent

The Amazon ECS Container Agent is a component of Amazon Elastic Container Service
(Amazon ECS) and is responsible for managing containers on behalf of Amazon ECS.


The best source of information on running this software is the
Amazon ECS documentation.

Please note that from Agent version 1.20.0, Minimum required Docker version is 1.9.0, corresponding to Docker API version 1.21. For more information, please visit Amazon ECS Container Agent Versions.

On the Amazon Linux AMI

On the Amazon Linux AMI, we provide an installable RPM which can be used via
sudo yum install ecs-init && sudo start ecs. This is the recommended way to run it in this environment.

On Other Linux AMIs

The Amazon ECS Container Agent may also be run in a Docker container on an EC2 instance with a recent Docker version
installed. A Docker image is available in our
Docker Hub Repository.

$ # Set up directories the agent uses
$ mkdir -p /var/log/ecs /etc/ecs /var/lib/ecs/data
$ touch /etc/ecs/ecs.config
$ # Set up necessary rules to enable IAM roles for tasks
$ sysctl -w net.ipv4.conf.all.route_localnet=1
$ iptables -t nat -A PREROUTING -p tcp -d --dport 80 -j DNAT --to-destination
$ iptables -t nat -A OUTPUT -d -p tcp -m tcp --dport 80 -j REDIRECT --to-ports 51679
$ # Run the agent
$ docker run --name ecs-agent \
--detach=true \
--restart=on-failure:10 \
--volume=/var/run/docker.sock:/var/run/docker.sock \
--volume=/var/log/ecs:/log \
--volume=/var/lib/ecs/data:/data \
--net=host \
--env-file=/etc/ecs/ecs.config \
--env=ECS_LOGFILE=/log/ecs-agent.log \
--env=ECS_DATADIR=/data/ \

On Other Linux AMIs when awsvpc networking mode is enabled

For the AWS VPC networking mode, ECS agent requires CNI plugin and dhclient to be available. ECS also needs the ecs-init to run as part of its startup.
The following is an example of docker run configuration for running ecs-agent with Task ENI enabled. Note that ECS agent currently only supports cgroupfs for cgroup driver.
$ # Run the agent
$ /usr/bin/docker run --name ecs-agent \
--init \
--restart=on-failure:10 \
--volume=/var/run:/var/run \
--volume=/var/log/ecs/:/log:Z \
--volume=/var/lib/ecs/data:/data:Z \
--volume=/etc/ecs:/etc/ecs \
--volume=/sbin:/host/sbin \
--volume=/lib:/lib \
--volume=/lib64:/lib64 \
--volume=/usr/lib:/usr/lib \
--volume=/usr/lib64:/usr/lib64 \
--volume=/proc:/host/proc \
--volume=/sys/fs/cgroup:/sys/fs/cgroup \
--net=host \
--env-file=/etc/ecs/ecs.config \
--cap-add=sys_admin \
--cap-add=net_admin \
--env ECS_ENABLE_TASK_ENI=true \
--env ECS_DATADIR=/data \
--env ECS_LOGFILE=/log/ecs-agent.log \
--env ECS_AVAILABLE_LOGGING_DRIVERS='["json-file","awslogs","syslog","none"]' \
--env ECS_LOGLEVEL=info \
--detach \

See also the Advanced Usage section below.

On the ECS Optimized Windows AMI

ECS Optimized Windows AMI ships with a pre-installed PowerShell module called ECSTools to install, configure, and run the ECS Agent as a Windows service.
To install the service, you can run the following PowerShell commands on an EC2 instance. To launch into another cluster instead of windows, replace the 'windows' in the script below with the name of your cluster.

PS C:\> Import-Module ECSTools
PS C:\> # The -EnableTaskIAMRole option is required to enable IAM roles for tasks.
PS C:\> Initialize-ECSAgent -Cluster 'windows' -EnableTaskIAMRole

Downloading Different Version of ECS Agent

To download different version of ECS Agent, you can do the following:

PS C:\> # use agentVersion = "latest" for the latest available agent version
PS C:\> $agentVersion = "v1.20.4"
PS C:\> Initialize-ECSAgent -Cluster 'windows' -EnableTaskIAMRole -Version $agentVersion

Advanced Usage

The Amazon ECS Container Agent supports a number of configuration options, most of which should be set through
environment variables.

Environment Variables

The table below provides an overview of optional environment variables that can be used to configure the ECS agent. See
the Amazon ECS developer guide for
additional details on each available environment variable.

| Environment Key | Example Value(s) | Description | Default value on Linux | Default value on Windows |
| ECS_CLUSTER | clusterName | The cluster this agent should check into. | default | default |
| ECS_RESERVED_PORTS | [22, 80, 5000, 8080] | An array of ports that should be marked as unavailable for scheduling on this container instance. | [22, 2375, 2376, 51678, 51679] | [53, 135, 139, 445, 2375, 2376, 3389, 5985, 5986, 51678, 51679]
| ECS_RESERVED_PORTS_UDP | [53, 123] | An array of UDP ports that should be marked as unavailable for scheduling on this container instance. | [] | [] |
| ECS_ENGINE_AUTH_TYPE | "docker" | "dockercfg" | The type of auth data that is stored in the ECS_ENGINE_AUTH_DATA key. | | |
| ECS_ENGINE_AUTH_DATA | See the dockerauth documentation | Docker auth data formatted as defined by ECS_ENGINE_AUTH_TYPE. | | |
| AWS_DEFAULT_REGION | <us-west-2>|<us-east-1>|… | The region to be used in API requests as well as to infer the correct backend host. | Taken from Amazon EC2 instance metadata. | Taken from Amazon EC2 instance metadata. |
| AWS_ACCESS_KEY_ID | AKIDEXAMPLE | The access key used by the agent for all calls. | Taken from Amazon EC2 instance metadata. | Taken from Amazon EC2 instance metadata. |
| AWS_SECRET_ACCESS_KEY | EXAMPLEKEY | The secret key used by the agent for all calls. | Taken from Amazon EC2 instance metadata. | Taken from Amazon EC2 instance metadata. |
| AWS_SESSION_TOKEN | | The session token used for temporary credentials. | Taken from Amazon EC2 instance metadata. | Taken from Amazon EC2 instance metadata. |
| DOCKER_HOST | unix:///var/run/docker.sock | Used to create a connection to the Docker daemon; behaves similarly to this environment variable as used by the Docker client. | unix:///var/run/docker.sock | npipe:////./pipe/docker_engine |
| ECS_LOGLEVEL | <crit> | <error> | <warn> | <info> | <debug> | The level of detail to be logged. | info | info |
| ECS_LOGLEVEL_ON_INSTANCE | <none> | <crit> | <error> | <warn> | <info> | <debug> | Can be used to override ECS_LOGLEVEL and set a level of detail that should be logged in the on-instance log file, separate from the level that is logged in the logging driver. If a logging driver is explicitly set, on-instance logs are turned off by default, but can be turned back on with this variable. | none if ECS_LOG_DRIVER is explicitly set to a non-empty value; otherwise the same value as ECS_LOGLEVEL | none if ECS_LOG_DRIVER is explicitly set to a non-empty value; otherwise the same value as ECS_LOGLEVEL |
| ECS_LOGFILE | /ecs-agent.log | The location where logs should be written. Log level is controlled by ECS_LOGLEVEL. | blank | blank |
| ECS_CHECKPOINT | <true | false> | Whether to checkpoint state to the DATADIR specified below. | true if ECS_DATADIR is explicitly set to a non-empty value; false otherwise | true if ECS_DATADIR is explicitly set to a non-empty value; false otherwise |
| ECS_DATADIR | /data/ | The container path where state is checkpointed for use across agent restarts. Note that on Linux, when you specify this, you will need to make sure that the Agent container has a bind mount of $ECS_HOST_DATA_DIR/data:$ECS_DATADIR with the corresponding values of ECS_HOST_DATA_DIR and ECS_DATADIR. | /data/ | C:\ProgramData\Amazon\ECS\data
| ECS_UPDATES_ENABLED | <true | false> | Whether to exit for an updater to apply updates when requested. | false | false |
| ECS_DISABLE_METRICS | <true | false> | Whether to disable metrics gathering for tasks. | false | true |
| ECS_POLL_METRICS | <true | false> | Whether to poll or stream when gathering metrics for tasks. Setting this value to true can help reduce the CPU usage of dockerd and containerd on the ECS container instance. See also ECS_POLL_METRICS_WAIT_DURATION for setting the poll interval. | false | false |
| ECS_POLLING_METRICS_WAIT_DURATION | 10s | Time to wait between polling for metrics for a task. Not used when ECS_POLL_METRICS is false. Maximum value is 20s and minimum value is 5s. If user sets above maximum it will be set to max, and if below minimum it will be set to min. | 10s | 10s |
| ECS_PULL_DEPENDENT_CONTAINERS_UPFRONT | <true | false> | Whether to pull images for containers with dependencies before the dependsOn condition has been satisfied. | false | false |
| ECS_RESERVED_MEMORY | 32 | Memory, in MiB, to reserve for use by things other than containers managed by Amazon ECS. | 0 | 0 |
| ECS_AVAILABLE_LOGGING_DRIVERS | ["awslogs","fluentd","gelf","json-file","journald","logentries","splunk","syslog"] | Which logging drivers are available on the container instance. | ["json-file","none"] | ["json-file","none"] |
| ECS_DISABLE_PRIVILEGED | true | Whether launching privileged containers is disabled on the container instance. | false | false |
| ECS_SELINUX_CAPABLE | true | Whether SELinux is available on the container instance. | false | false |
| ECS_APPARMOR_CAPABLE | true | Whether AppArmor is available on the container instance. | false | false |
| ECS_ENGINE_TASK_CLEANUP_WAIT_DURATION | 10m | Default time to wait to delete containers for a stopped task (see also ECS_ENGINE_TASK_CLEANUP_WAIT_DURATION_JITTER). If set to less than 1 minute, the value is ignored. | 3h | 3h |
| ECS_ENGINE_TASK_CLEANUP_WAIT_DURATION_JITTER | 1h | Jitter value for the task engine cleanup wait duration. When specified, the actual cleanup wait duration time for each task will be the duration specified in ECS_ENGINE_TASK_CLEANUP_WAIT_DURATION plus a random duration between 0 and the jitter duration. | blank | blank |
| ECS_CONTAINER_STOP_TIMEOUT | 10m | Instance scoped configuration for time to wait for the container to exit normally before being forcibly killed. | 30s | 30s |
| ECS_CONTAINER_START_TIMEOUT | 10m | Timeout before giving up on starting a container. | 3m | 8m |
| ECS_CONTAINER_CREATE_TIMEOUT | 10m | Timeout before giving up on creating a container. Minimum value is 1m. If user sets a value below minimum it will be set to min. | 4m | 4m |
| ECS_ENABLE_TASK_IAM_ROLE | true | Whether to enable IAM Roles for Tasks on the Container Instance | false | false |
| ECS_ENABLE_TASK_IAM_ROLE_NETWORK_HOST | true | Whether to enable IAM Roles for Tasks when launched with host network mode on the Container Instance | false | false |
| ECS_DISABLE_IMAGE_CLEANUP | true | Whether to disable automated image cleanup for the ECS Agent. | false | false |
| ECS_IMAGE_CLEANUP_INTERVAL | 30m | The time interval between automated image cleanup cycles. If set to less than 10 minutes, the value is ignored. | 30m | 30m |
| ECS_IMAGE_MINIMUM_CLEANUP_AGE | 30m | The minimum time interval between when an image is pulled and when it can be considered for automated image cleanup. | 1h | 1h |
| NON_ECS_IMAGE_MINIMUM_CLEANUP_AGE | 30m | The minimum time interval between when a non ECS image is created and when it can be considered for automated image cleanup. | 1h | 1h |
| ECS_NUM_IMAGES_DELETE_PER_CYCLE | 5 | The maximum number of images to delete in a single automated image cleanup cycle. If set to less than 1, the value is ignored. | 5 | 5 |
| ECS_IMAGE_PULL_BEHAVIOR | <default | always | once | prefer-cached > | The behavior used to customize the pull image process. If default is specified, the image will be pulled remotely, if the pull fails then the cached image in the instance will be used. If always is specified, the image will be pulled remotely, if the pull fails then the task will fail. If once is specified, the image will be pulled remotely if it has not been pulled before or if the image was removed by image cleanup, otherwise the cached image in the instance will be used. If prefer-cached is specified, the image will be pulled remotely if there is no cached image, otherwise the cached image in the instance will be used. | default | default |
| ECS_IMAGE_PULL_INACTIVITY_TIMEOUT | 1m | The time to wait after docker pulls complete waiting for extraction of a container. Useful for tuning large Windows containers. | 1m | 3m |
| ECS_IMAGE_PULL_TIMEOUT | 1h | The time to wait for pulling docker image. | 2h | 2h |
| ECS_INSTANCE_ATTRIBUTES | {"stack": "prod"} | These attributes take effect only during initial registration. After the agent has joined an ECS cluster, use the PutAttributes API action to add additional attributes. For more information, see Amazon ECS Container Agent Configuration in the Amazon ECS Developer Guide.| {} | {} |
| ECS_ENABLE_TASK_ENI | false | Whether to enable task networking for task to be launched with its own network interface | false | Not applicable |
| ECS_ENABLE_HIGH_DENSITY_ENI | false | Whether to enable high density eni feature when using task networking | true | Not applicable |
| ECS_CNI_PLUGINS_PATH | /ecs/cni | The path where the cni binary file is located | /amazon-ecs-cni-plugins | Not applicable |
| ECS_AWSVPC_BLOCK_IMDS | true | Whether to block access to Instance Metadata for Tasks started with awsvpc network mode | false | Not applicable |
| ECS_AWSVPC_ADDITIONAL_LOCAL_ROUTES | [""] | In awsvpc network mode, traffic to these prefixes will be routed via the host bridge instead of the task ENI | [] | Not applicable |
| ECS_ENABLE_CONTAINER_METADATA | true | When true, the agent will create a file describing the container's metadata and the file can be located and consumed by using the container enviornment variable $ECS_CONTAINER_METADATA_FILE | false | false |
| ECS_HOST_DATA_DIR | /var/lib/ecs | The source directory on the host from which ECS_DATADIR is mounted. We use this to determine the source mount path for container metadata files in the case the ECS Agent is running as a container. We do not use this value in Windows because the ECS Agent is not running as container in Windows. On Linux, note that when you specify this, you will need to make sure that the Agent container has a bind mount of $ECS_HOST_DATA_DIR/data:$ECS_DATADIR with the corresponding values of ECS_HOST_DATA_DIR and ECS_DATADIR. | /var/lib/ecs | Not used |
| ECS_ENABLE_TASK_CPU_MEM_LIMIT | true | Whether to enable task-level cpu and memory limits | true | false |
| ECS_CGROUP_PATH | /sys/fs/cgroup | The root cgroup path that is expected by the ECS agent. This is the path that accessible from the agent mount. | /sys/fs/cgroup | Not applicable |
| ECS_CGROUP_CPU_PERIOD | 10ms | CGroups CPU period for task level limits. This value should be between 8ms to 100ms | 100ms | Not applicable |
| ECS_AGENT_HEALTHCHECK_HOST | localhost | Override for the ecs-agent container's healthcheck localhost ip address| localhost | localhost |
| ECS_ENABLE_CPU_UNBOUNDED_WINDOWS_WORKAROUND | true | When true, ECS will allow CPU unbounded(CPU=0) tasks to run along with CPU bounded tasks in Windows. | Not applicable | false |
| ECS_ENABLE_MEMORY_UNBOUNDED_WINDOWS_WORKAROUND | true | When true, ECS will ignore the memory reservation parameter (soft limit) to run along with memory bounded tasks in Windows. To run a memory unbounded task, omit the memory hard limit and set any memory reservation, it will be ignored. | Not applicable | false |
| ECS_TASK_METADATA_RPS_LIMIT | 100,150 | Comma separated integer values for steady state and burst throttle limits for task metadata endpoint | 40,60 | 40,60 |
| ECS_SHARED_VOLUME_MATCH_FULL_CONFIG | true | When true, ECS Agent will compare name, driver options, and labels to make sure volumes are identical. When false, Agent will short circuit shared volume comparison if the names match. This is the default Docker behavior. If a volume is shared across instances, this should be set to false. | false | false|
| ECS_CONTAINER_INSTANCE_PROPAGATE_TAGS_FROM | ec2_instance | If ec2_instance is specified, existing tags defined on the container instance will be registered to Amazon ECS and will be discoverable using the ListTagsForResource API. Using this requires that the IAM role associated with the container instance have the ec2:DescribeTags action allowed. | none | none |
| ECS_CONTAINER_INSTANCE_TAGS | {"tag_key": "tag_val"} | The metadata that you apply to the container instance to help you categorize and organize them. Each tag consists of a key and an optional value, both of which you define. Tag keys can have a maximum character length of 128 characters, and tag values can have a maximum length of 256 characters. If tags also exist on your container instance that are propagated using the ECS_CONTAINER_INSTANCE_PROPAGATE_TAGS_FROM parameter, those tags will be overwritten by the tags specified using ECS_CONTAINER_INSTANCE_TAGS. | {} | {} |
| ECS_ENABLE_UNTRACKED_IMAGE_CLEANUP | true | Whether to allow the ECS agent to delete containers and images that are not part of ECS tasks. | false | false |
| ECS_EXCLUDE_UNTRACKED_IMAGE | alpine:latest | Comma seperated list of imageName:tag of images that should not be deleted by the ECS agent if ECS_ENABLE_UNTRACKED_IMAGE_CLEANUP is enabled. | | |
| ECS_DISABLE_DOCKER_HEALTH_CHECK | false | Whether to disable the Docker Container health check for the ECS Agent. | false | false |
| ECS_NVIDIA_RUNTIME | nvidia | The Nvidia Runtime to be used to pass Nvidia GPU devices to containers. | nvidia | Not Applicable |
| ECS_ENABLE_SPOT_INSTANCE_DRAINING | true | Whether to enable Spot Instance draining for the container instance. If true, if the container instance receives a spot interruption notice, agent will set the instance's status to DRAINING, which gracefully shuts down and replaces all tasks running on the instance that are part of a service. It is recommended that this be set to true when using spot instances. | false | false |
| ECS_LOG_ROLLOVER_TYPE | size | hourly | Determines whether the container agent logfile will be rotated based on size or hourly. By default, the agent logfile is rotated each hour. | hourly | hourly |
| ECS_LOG_OUTPUT_FORMAT | logfmt | json | Determines the log output format. When the json format is used, each line in the log would be a structured JSON map. | logfmt | logfmt |
| ECS_LOG_MAX_FILE_SIZE_MB | 10 | When the ECS_LOG_ROLLOVER_TYPE variable is set to size, this variable determines the maximum size (in MB) the log file before it is rotated. If the rollover type is set to hourly then this variable is ignored. | 10 | 10 |
| ECS_LOG_MAX_ROLL_COUNT | 24 | Determines the number of rotated log files to keep. Older log files are deleted once this limit is reached. | 24 | 24 |
| ECS_LOG_DRIVER | awslogs | fluentd | gelf | json-file | journald | logentries | syslog | splunk | The logging driver to be used by the Agent container. | json-file | Not applicable |
| ECS_LOG_OPTS | {"option":"value"} | The options for configuring the logging driver set in ECS_LOG_DRIVER. | {} | Not applicable |
| ECS_ENABLE_AWSLOGS_EXECUTIONROLE_OVERRIDE | true | Whether to enable awslogs log driver to authenticate via credentials of task execution IAM role. Needs to be true if you want to use awslogs log driver in a task that has task execution IAM role specified. When using the ecs-init RPM with version equal or later than V1.16.0-1, this env is set to true by default. | false | false |
| ECS_FSX_WINDOWS_FILE_SERVER_SUPPORTED | true | Whether FSx for Windows File Server volume type is supported on the container instance. This variable is only supported on agent versions 1.47.0 and later. | false | true |
| ECS_ENABLE_RUNTIME_STATS | true | Determines if pprof is enabled for the agent. If enabled, the different profiles can be accessed through the agent's introspection port (e.g. curl http://localhost:51678/debug/pprof/heap > heap.pprof). In addition, agent's runtime stats are logged to /var/log/ecs/runtime-stats.log file. | false | false |


When you run the Amazon ECS Container Agent in production, its datadir should be persisted between runs of the Docker
container. If this data is not persisted, the agent registers a new container instance ARN on each launch and is not
able to update the state of tasks it previously ran.


The agent also supports the following flags:

Building and Running from Source

Running the Amazon ECS Container Agent outside of Amazon EC2 is not supported.

Docker Image (on Linux)

The Amazon ECS Container Agent may be built by typing make with the Docker
daemon (v1.5.0) running.

This produces an image tagged amazon/ecs-container-agent:make that
you may run as described above.

Standalone (on Linux)

The Amazon ECS Container Agent may also be run outside of a Docker container as a Go binary. This is not recommended
for production on Linux, but it can be useful for development or easier integration with your local Go tools.

The following commands run the agent outside of Docker:

make gobuild

Make Targets (on Linux)

The following targets are available. Each may be run with make <target>.

| Make Target | Description |
| release | (Default) Builds the agent within a Docker container and and packages it into a scratch-based image |
| gobuild | Runs a normal go build of the agent and stores the binary in ./out/amazon-ecs-agent |
| static | Runs go build to produce a static binary in ./out/amazon-ecs-agent |
| test | Runs all unit tests using go test |
| test-in-docker | Runs all tests inside a Docker container |
| run-integ-tests | Runs all integration tests in the engine and stats packages |
| clean | Removes build artifacts. Note: this does not remove Docker images |

Standalone (on Windows)

The Amazon ECS Container Agent may be built by invoking scripts\build_agent.ps1

Scripts (on Windows)

The following scripts are available to help develop the Amazon ECS Container Agent on Windows:


Contributions and feedback are welcome! Proposals and pull requests will be considered and responded to. For more
information, see the file.

If you have a bug/and issue around the behavior of the ECS agent, please open it here.

If you have a feature request, please open it over at the AWS Containers Roadmap.

Amazon Web Services does not currently provide support for modified copies of this software.

Security disclosures

If you think you’ve found a potential security issue, please do not post it in the Issues. Instead, please follow the instructions here or email AWS security directly.


The Amazon ECS Container Agent is licensed under the Apache 2.0 License.