ECS config discovery: "Router defined multiple times with different configurations"

Traefik Traefik v3 (latest)
1

Hi,

I'm deploying Traefik to AWS ECS to proxy traffic to ECS tasks that it finds through the ECS configuration discovery. It works perfectly when the task is running in a stable state, but during deployments (when there is >1 task instance) I get "Router defined multiple times with different configurations" errors.

The errors look like this

	
{"level":"error","providerName":"ecs","routerName": "router_name","configuration":["service-app_name-HASH_1","service-app_name-HASH_2"],"time":"2025-02-10T11:33:23Z","message":"Router defined multiple times with different configurations"}

I believe HASH_1 and HASH_2 are related to the two ECS tasks, but I can't be sure of that because they don't match up with any identifiers in the console.

The docker labels I attach to my running container looks like this

      dockerLabels = {
        "traefik.http.routers.app_name-${var.pull_request_number}.rule" : "Host(`pr-${var.pull_request_number}.app.domain`)",
        "traefik.http.routers.app_name-${var.pull_request_number}.service" : "app_name-${var.pull_request_number}",
        "traefik.http.services.app_name-${var.pull_request_number}.loadbalancer.server.port" : "3000",
        "traefik.http.services.app_name-${var.pull_request_number}.loadbalancer.healthcheck.path" : "/up",
        "traefik.enable" : "true"
      },

Is this just a known thing with ECS configuration discovery, or do I need to make my router and service names totally unique, perhaps by including a timestamp? The ultimate goal is to achieve a rolling deploy and as minimal downtime as possible.