As AWS still seems to be the most popular k8s cluster provider for installing Workflow, seemingly even for dev clusters, perhaps the annotation mentioned in https://deis.com/docs/workflow/managing-workflow/configuring-load-balancers/#idle-connection-timeouts should be added to the default router service template so that operators don't bump into issues stemming from the default load balancer timeout of 60s, as seems to be the frequent case currently.
Agree. I actually implemented that feature upstream in k8s with the intention of making this same modification to the chart... but then I forgot about it.
As AWS still seems to be the most popular k8s cluster provider for installing Workflow, seemingly even for dev clusters, perhaps the annotation mentioned in https://deis.com/docs/workflow/managing-workflow/configuring-load-balancers/#idle-connection-timeouts should be added to the default router service template so that operators don't bump into issues stemming from the default load balancer timeout of 60s, as seems to be the frequent case currently.
Thoughts?