HealthCheck - Amazon Elastic Container Service
Services or capabilities described in Amazon Web Services documentation might vary by Region. To see the differences applicable to the China Regions, see Getting Started with Amazon Web Services in China (PDF).

HealthCheck

An object representing a container health check. Health check parameters that are specified in a container definition override any Docker health checks that exist in the container image (such as those specified in a parent image or from the image's Dockerfile). This configuration maps to the HEALTHCHECK parameter of docker run.

Note

The Amazon ECS container agent only monitors and reports on the health checks specified in the task definition. Amazon ECS does not monitor Docker health checks that are embedded in a container image and not specified in the container definition. Health check parameters that are specified in a container definition override any Docker health checks that exist in the container image.

You can view the health status of both individual containers and a task with the DescribeTasks API operation or when viewing the task details in the console.

The health check is designed to make sure that your containers survive agent restarts, upgrades, or temporary unavailability.

Amazon ECS performs health checks on containers with the default that launched the container instance or the task.

The following describes the possible healthStatus values for a container:

  • HEALTHY-The container health check has passed successfully.

  • UNHEALTHY-The container health check has failed.

  • UNKNOWN-The container health check is being evaluated, there's no container health check defined, or Amazon ECS doesn't have the health status of the container.

The following describes the possible healthStatus values based on the container health checker status of essential containers in the task with the following priority order (high to low):

  • UNHEALTHY-One or more essential containers have failed their health check.

  • UNKNOWN-Any essential container running within the task is in an UNKNOWN state and no other essential containers have an UNHEALTHY state.

  • HEALTHY-All essential containers within the task have passed their health checks.

Consider the following task health example with 2 containers.

  • If Container1 is UNHEALTHY and Container2 is UNKNOWN, the task health is UNHEALTHY.

  • If Container1 is UNHEALTHY and Container2 is HEALTHY, the task health is UNHEALTHY.

  • If Container1 is HEALTHY and Container2 is UNKNOWN, the task health is UNKNOWN.

  • If Container1 is HEALTHY and Container2 is HEALTHY, the task health is HEALTHY.

Consider the following task health example with 3 containers.

  • If Container1 is UNHEALTHY and Container2 is UNKNOWN, and Container3 is UNKNOWN, the task health is UNHEALTHY.

  • If Container1 is UNHEALTHY and Container2 is UNKNOWN, and Container3 is HEALTHY, the task health is UNHEALTHY.

  • If Container1 is UNHEALTHY and Container2 is HEALTHY, and Container3 is HEALTHY, the task health is UNHEALTHY.

  • If Container1 is HEALTHY and Container2 is UNKNOWN, and Container3 is HEALTHY, the task health is UNKNOWN.

  • If Container1 is HEALTHY and Container2 is UNKNOWN, and Container3 is UNKNOWN, the task health is UNKNOWN.

  • If Container1 is HEALTHY and Container2 is HEALTHY, and Container3 is HEALTHY, the task health is HEALTHY.

If a task is run manually, and not as part of a service, the task will continue its lifecycle regardless of its health status. For tasks that are part of a service, if the task reports as unhealthy then the task will be stopped and the service scheduler will replace it.

The following are notes about container health check support:

  • If the Amazon ECS container agent becomes disconnected from the Amazon ECS service, this won't cause a container to transition to an UNHEALTHY status. This is by design, to ensure that containers remain running during agent restarts or temporary unavailability. The health check status is the "last heard from" response from the Amazon ECS agent, so if the container was considered HEALTHY prior to the disconnect, that status will remain until the agent reconnects and another health check occurs. There are no assumptions made about the status of the container health checks.

  • Container health checks require version 1.17.0 or greater of the Amazon ECS container agent. For more information, see Updating the Amazon ECS container agent.

  • Container health checks are supported for Fargate tasks if you're using platform version 1.1.0 or greater. For more information, see Amazon Fargate platform versions.

  • Container health checks aren't supported for tasks that are part of a service that's configured to use a Classic Load Balancer.

Contents

command

A string array representing the command that the container runs to determine if it is healthy. The string array must start with CMD to run the command arguments directly, or CMD-SHELL to run the command with the container's default shell.

When you use the Amazon Management Console JSON panel, the Amazon Command Line Interface, or the APIs, enclose the list of commands in double quotes and brackets.

[ "CMD-SHELL", "curl -f http://localhost/ || exit 1" ]

You don't include the double quotes and brackets when you use the Amazon Management Console.

CMD-SHELL, curl -f http://localhost/ || exit 1

An exit code of 0 indicates success, and non-zero exit code indicates failure. For more information, see HealthCheck in the docker container create command.

Type: Array of strings

Required: Yes

interval

The time period in seconds between each health check execution. You may specify between 5 and 300 seconds. The default value is 30 seconds.

Type: Integer

Required: No

retries

The number of times to retry a failed health check before the container is considered unhealthy. You may specify between 1 and 10 retries. The default value is 3.

Type: Integer

Required: No

startPeriod

The optional grace period to provide containers time to bootstrap before failed health checks count towards the maximum number of retries. You can specify between 0 and 300 seconds. By default, the startPeriod is off.

Note

If a health check succeeds within the startPeriod, then the container is considered healthy and any subsequent failures count toward the maximum number of retries.

Type: Integer

Required: No

timeout

The time period in seconds to wait for a health check to succeed before it is considered a failure. You may specify between 2 and 60 seconds. The default value is 5.

Type: Integer

Required: No

See Also

For more information about using this API in one of the language-specific Amazon SDKs, see the following: