Hello everybody,
we just started using Traefik 2.3 RC and we like it a lot.
After using docker as provider, we wanted to switch to ECS as provider.
Our setup in AWS:
We set up an NLB which forwards traffic to Traefik(no encryption used) and then Traefik is supposed to get all information about the services running in the backend from ECS. Traefik is running as its own service in ECS. On the same cluster we are running our backend service, again as its own service in ECS.
The command for the Traefik service container is:
Command ["--log.level=DEBUG","--api.dashboard=true","--api.insecure=true","--ping=true","--providers.ecs.autoDiscoverClusters=false","--providers.ecs.clusters=mycluster","--providers.ecs.region=us-east-1","--providers.ecs.exposedByDefault=false","--entrypoints.web.address=:80"]
Sending a request to the NLB, we see an error in the console which states:
Unable to list tasks" providerName=ecs
Provider connection error RequestError: send request failed\caused by : Post "https://ecs.us-eas-1.amazonaws.com/\": dial : i/o timeout, retrying .....
Note: I replaced the real IP with someIp.
We do not have a VPC endpoint for ECS and we are thinking that this could be the reason why Traefik cannot connect to it.
What else could be the reason?
In case it is the missing VPC endpoint and assuming we cannot get a VPC endpoint set up in a short time, what other options do we have to make this connection between Traefik and ECS work?
Thanks