ecs-cli

Run ad-hoc containers on ECS.

This does not aim to replace the offical AWS ECS CLI, but complement it. The official CLI doesn't support docker run like commands.

Goals

This CLI is designed to closely resemble the docker run command for ECS. This will allow users to run any container against a cluster without previously having to define a task definition. It will provision the necessary AWS resources like the task definition and CloudWatch log group, start your container, and stream the logs back to stdout.

• replicate docker run arguments
• provision the necessary AWS resources to run a container
  • allow custom task-definition name
• the exit code from the container is mirrored by the CLI
• capture a SIGKILL event (Ctrl+c) and stop the remote container
• replicate docker exec arguments using SSM agent

Installation

Download the latest release for your OS and extract to your PATH.

Example

➜  ~ ecs run --cluster ops bash ping -c 5 google.com
Creating task definition
Running task definition: arn:aws:ecs:us-east-1:000000000000:task-definition/ephemeral-task-from-ecs-cli:1
Streaming from Cloudwatch Logs
https://console.aws.amazon.com/ecs/home?#/clusters/ops/tasks/00000000000000000000000000000000/details
Container is starting on EC2 instance i-00000000000000000 (10.100.1.25).
2019-01-19 20:25:08 -0600 CST	PING google.com (172.217.7.238): 56 data bytes
2019-01-19 20:25:08 -0600 CST	64 bytes from 172.217.7.238: seq=0 ttl=46 time=2.020 ms
2019-01-19 20:25:09 -0600 CST	64 bytes from 172.217.7.238: seq=1 ttl=46 time=1.303 ms
2019-01-19 20:25:10 -0600 CST	64 bytes from 172.217.7.238: seq=2 ttl=46 time=1.263 ms
2019-01-19 20:25:11 -0600 CST	64 bytes from 172.217.7.238: seq=3 ttl=46 time=1.230 ms
2019-01-19 20:25:12 -0600 CST	64 bytes from 172.217.7.238: seq=4 ttl=46 time=1.313 ms
2019-01-19 20:25:12 -0600 CST	--- google.com ping statistics ---
2019-01-19 20:25:12 -0600 CST	5 packets transmitted, 5 packets received, 0% packet loss
2019-01-19 20:25:12 -0600 CST	round-trip min/avg/max = 1.230/1.425/2.020 ms
Task arn:aws:ecs:us-east-1:000000000000:task/ops/00000000000000000000000000000000 has stopped (exit code 0):
	Essential container in task exited
All containers have exited

Usage

Usage:
  ecs run [flags]

Flags:
      --cluster string           ECS cluster
  -c, --count int                Spawn n tasks (default 1)
      --cpu-reservation int      CPU reservation (default 1024)
  -d, --detach                   Run the task in the background
  -e, --env stringArray          Set environment variables
      --execution-role string    Execution role ARN (required for Fargate)
      --fargate                  Launch in Fargate
  -h, --help                     help for run
  -m, --memory int               Memory limit
      --memory-reservation int   Memory reservation (default 2048)
  -n, --name string              Assign a name to the task (default "ecs-cli-app")
  -p, --publish stringArray      Publish a container's port(s) to the host
      --subnet stringArray       Subnet(s) where task should run
  -v, --volume stringArray       Map volume to ECS Container Instance