Help with AWS ECS Deployment

[Nirosh-cartrabbit]

Hii I'm trying to deploy my Dagster project to ECS but when i trying to deploy them in the cluster as service my tasks is not running it just draining every time. My task definition consists of the three container like told in docs, when i create my service and add my task to it my Dagster dosen't seems to get up and running.

my dagster yaml file is

storage:
  postgres:
    postgres_url:
      env: DAGSTER_POSTGRES_DB

telemetry:
  enabled: false

run_monitoring:
  enabled: true
  start_timeout_seconds: 180
  cancel_timeout_seconds: 180
  poll_interval_seconds: 120
  free_slots_after_run_end_seconds: 2

run_retries:
  enabled: true
  max_retries: 5
  retry_on_asset_or_op_failure: true

compute_logs:
  module: dagster_aws.s3.compute_log_manager
  class: S3ComputeLogManager
  config:
    bucket: beta-tasks-dev
    region: us-east-2

Do i need to have launcher type as ecs ah. how does that work can any one explain and help me with this ?

[garethbrickman]

Could you link to the doc you are using for reference?

[Nirosh-cartrabbit]

This is the docs im referring to https://docs.dagster.io/deployment/guides/aws

my doubt is mainly do i really need to change my run launcher to ECS ah

run_launcher:
  module: "dagster_aws.ecs"
  class: "EcsRunLauncher"
  config:
    task_definition: "arn:aws:ecs:us-east-2:2323:task-definition/tasksdev:19"
    container_name: "betatask-grpc"
  run_resources:
    cpu: "256"
    memory: "512" # In MiB
    ephemeral_storage: 20 # In GiB

or can i use deafult run launcher ah ?

[garethbrickman]

According to the Deploying in ECS doc: The Dagster instance uses the EcsRunLauncher to launch each run in its own ECS task.

[Nirosh-cartrabbit]

Thank you for the reply. But it works fine in normal run launcher itself. ECS tasks takes time to up and provision before that my run monitoring makes it failed. I cant understand why it is necessary to use ECS launcher, can u help me understand or navigate me where to learn about this??

problem i have with my current deployment now is this error message

MONITORING daemon is still sending heartbeats. You likely have multiple daemon processes running at once, which is not supported

can u help me with this ?