Zone aware routing
We use the following terminology:
Originating/Upstream cluster: Envoy routes requests from an originating cluster to an upstream cluster.
Local zone: The same zone that contains a subset of hosts in both the originating and upstream clusters.
Zone aware routing: Best effort routing of requests to an upstream cluster host in the local zone.
In deployments where hosts in originating and upstream clusters belong to different zones Envoy performs zone aware routing. There are several preconditions before zone aware routing can be performed:
Both originating and upstream cluster are not in panic mode.
Zone aware routing is enabled.
The upstream cluster has enough hosts. See here for more information.
The purpose of zone aware routing is to send as much traffic to the local zone in the upstream cluster as possible while roughly maintaining the same number of requests per second across all upstream hosts (depending on load balancing policy).
Envoy tries to push as much traffic as possible to the local upstream zone as long as roughly the same number of requests per host in the upstream cluster are maintained. The decision of whether Envoy routes to the local zone or performs cross zone routing depends on the percentage of healthy hosts in the originating cluster and upstream cluster in the local zone. There are two cases with regard to percentage relations in the local zone between originating and upstream clusters:
The originating cluster local zone percentage is greater than the one in the upstream cluster. In this case we cannot route all requests from the local zone of the originating cluster to the local zone of the upstream cluster because that will lead to request imbalance across all upstream hosts. Instead, Envoy calculates the percentage of requests that can be routed directly to the local zone of the upstream cluster. The rest of the requests are routed cross zone. The specific zone is selected based on the residual capacity of the zone (that zone will get some local zone traffic and may have additional capacity Envoy can use for cross zone traffic).
The originating cluster local zone percentage is smaller than the one in upstream cluster. In this case the local zone of the upstream cluster can get all of the requests from the local zone of the originating cluster and also have some space to allow traffic from other zones in the originating cluster (if needed).
Note that when using multiple priorities, zone aware routing is currently only supported for P=0.