Open abespalko opened 1 year ago
Thanks for opening this bug @abespalko. SC Agent becoming unhealthy here as indicated by the logs is due to failure to reach another component on the instance called as Relay agent that is in the path to connection with our management service.
We would need to investigate whether there was an issue on the instance agents, could you collect logs from the instance and see if something is evident there? You can use this instructions to collect all the logs: https://docs.aws.amazon.com/AmazonECS/latest/developerguide/ecs-logs-collector.html. You can share them with us using premium support or via email at ecs-service-connect-agent-external@amazon.com.
We also noticed that your host has 1 vCPU and running 7 containers so, their maybe contention there. Please refer this doc for the recommended CPU & Memory allocation values: https://docs.aws.amazon.com/AmazonECS/latest/developerguide/service-connect-concepts.html#service-connect-concepts-proxy
Summary
ECS Service connect agent becomes unhealthy.
Description
I have ECS cluster with several Services running under EC2 launch type (2-3 t2.micro) and some of them use ServiceConnect feature of AWS ECS. However, from time to time services that are in bridge or awsvpc network are just restarted due to
ecs-service-connect-agent:interface-v1
sidecar container becomes UNHEALTHY. All subsequent deployments of Services to that instance are failing due to ecs service connect agent healthcheck. After that I have to restart EC2 instance (maybe docker restart would also help?).Backend:
Redis:
Frontend:
Expected Behavior
Container Image: ecs-service-connect-agent:interface-v1 Logs on success deployment:
Observed Behavior
Environment Details