You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
This is a feature request for Swarm Mode. There is a plan to support topology-aware task scheduling in 1.14 (moby/swarmkit#1512). This feature request is to take same model and ensure that task affinity is taken into consideration when resolving DNS queries for services. For example, if we deploy a stack made up of two services ( web and k/v store ) each with two replicas across two availability zones. It would be optimal for web_1 task to talk to kv_1 when both are deployed on same AZ and fall back to talk to kv_2 if kv_1 is down.
For routing mesh's ingress network, this feature can ensure that local tasks are prioritized when an external request comes in. Instead of current mode where requests for an exposed service can be forwarded to tasks running on other nodes even if there is a task running on the node that receives the request.
Description
This is a feature request for Swarm Mode. There is a plan to support topology-aware task scheduling in 1.14 (moby/swarmkit#1512). This feature request is to take same model and ensure that task affinity is taken into consideration when resolving DNS queries for services. For example, if we deploy a stack made up of two services ( web and k/v store ) each with two replicas across two availability zones. It would be optimal for
web_1task to talk tokv_1when both are deployed on same AZ and fall back to talk tokv_2ifkv_1is down.For routing mesh's
ingressnetwork, this feature can ensure that local tasks are prioritized when an external request comes in. Instead of current mode where requests for an exposed service can be forwarded to tasks running on other nodes even if there is a task running on the node that receives the request.