Kubernetes : Some Warnings in Production Mode

Hi, Noticed those logs on our Kubernetes SC server pod (3 Workers + 2 Brokers) :

1552651454358 - Worker 2 was respawned 1 1552651454751 - Origin: Worker (PID 1839) [Warning] SocketProtocolError: Client pong timed out at SCServerSocket._onSCClose (/usr/src/node_modules/socketcluster-server/scserversocket.js:249:17) at Timeout.<anonymous> (/usr/src/node_modules/socketcluster-server/scserversocket.js:194:10) at ontimeout (timers.js:498:11) at tryOnTimeout (timers.js:323:5) at Timer.listOnTimeout (timers.js:290:5) [ 'From : ', '044191', ', To :', 'M00823' ] { channel: 'ROOM_044191' }

Please can You explain the "respawned status" effect on pending messages ? Also Is that the cause of client pong timed out ?

PS: For the configuration I changed the service type NodePort as default in your file "socketcluster-service.yaml" to a normal ClusterIP. Little confusing when your already have Ingress to forward the requests.

Env vars configuration :
"env": [ { "name": "NODE_ENV", "value": "prod" }, { "name": "SCC_STATE_SERVER_HOST", "value": "scc-state" }, { "name": "SOCKETCLUSTER_WORKERS", "value": "3" }, { "name": "SOCKETCLUSTER_BROKERS", "value": "2" }, { "name": "SCC_INSTANCE_IP", "valueFrom": { "fieldRef": { "apiVersion": "v1", "fieldPath": "status.podIP" } } }, { "name": "ENV", "value": "prod" } ]

Start by reviewing socketcluster-server/scserversocket.js at the logged _onSCClose and timeout locations, then inspect socketcluster-service.yaml and the supplied worker, broker, and Kubernetes settings. Reproduce or trace the warning in production mode and document how worker respawning affects pending messages, whether it relates to the pong timeout, and how the ClusterIP setting interacts with Ingress.