[nginx-ingress] increase timeout for ELBs

[osterman]

what

• Increase connection timeout for helm with service.beta.kubernetes.io/aws-load-balancer-connection-idle-timeout: "3600"

why

• Helm operations can easily take longer than the default 60 second TTL
• Error: could not ping Tiller: rpc error: code = Unavailable desc = transport is closing

references

• "Error: Transport is closing" message when attempting to install helm/helm#3409 (comment)
• https://medium.com/kokster/how-to-setup-nginx-ingress-controller-on-aws-clusters-7bd244278509

[osterman]

3600 might be a bit extreme. We can start with 300 (5min); default is 60.