Multi-zone deployment

About

Kuma supports running your service mesh in multiple zones. It is even possible to run with a mix of Kubernetes and Universal zones. Your mesh environment can include multiple isolated service meshes (multi-tenancy), and workloads running in different regions, on different clouds, or in different datacenters. A zone can be a Kubernetes cluster, a VPC, or any other deployment you need to include in the same distributed mesh environment. The only condition is that all the data planes running within the zone must be able to connect to the other data planes in this same zone.

Multi-zone deployment - 图1

How it works

In Kuma, zones are abstracted away, meaning that your data plane proxies will find services anywhere they run. Therefore, you can make a service multi-zone by having data planes using the same kuma.io/service run in different zones, this can help you achieve automatic fail-over of services when a specific zone fails.

We will now explain how this works in details:

In the local zone: the zone ingress will receive all traffic coming from other zones and route it within this zone. The control-plane will update its local zone ingresses with a list of local services and the number of instances available, it will then synchronize it with the global control-plane.

The global control-plane will propagate the zone ingress resources and all policies to all other zones over Kuma Discovery Service (KDS), which is a protocol based on xDS.

In the remote zone, data plane proxies will either add new services or add endpoints to existing services which correspond to the remote zone-ingresses entries. Requests are then routed to either local instances of the service or to the remote zone ingress where instances of the service are running, which will proxy the request to the actual instance.

In the presence of a zone egress the traffic is routed through the local zone egress before being sent to the remote zone ingress.

When using transparent-proxy (enabled by default in Kubernetes), Kuma generates a VIP, a DNS entry with the format <kuma.io/service>.mesh, and will listen for traffic on port 80.

A zone ingress is not an API gateway. It is only used for cross-zone communication within a mesh. API gateways are supported in Kuma gateway mode and can be deployed in addition to zone ingresses.

For Kubernetes the kuma.io/service is automatically generated as explained in the data-plane on Kubernetes documentation.

For load-balancing the zone ingress endpoints are weighted with the number of instances running behind them (.i.e: a zone with 2 instances will receive twice more traffic than a zone with 1 instance), you can also favor local traffic with localityAware load-balancing.

The <kuma.io/service>.mesh:80> is a convention. Virtual Outbounds will enable you to expose hostname/port differently.

Components of a multi-zone deployment

A multi-zone deployment includes:

  • The global control plane:
    • Accept connections only from zone control planes.
    • Accept creation and changes to policies that will be applied to the data plane proxies.
    • Send policies down to zone control planes.
    • Send zone ingresses down to zone control plane.
    • Keep an inventory of all data plane proxies running in all zones (this is only done for observability but is not required for operations).
    • Reject connections from data plane proxies.
  • The zone control planes:
    • Accept connections from data plane proxies started within this zone.
    • Receive policy updates from the global control plane.
    • Send data plane proxies and zone ingress changes to the global control plane.
    • Compute and send configurations using XDS to the local data plane proxies.
    • Update list of services which exist in the zone in the zone ingress.
    • Reject policy changes that do not come from global.
  • The data plane proxies:
    • Connect to the local zone control plane.
    • Receive configurations using XDS from the local zone control plane.
    • Connect to other local data plane proxies.
    • Connect to zone ingresses for sending cross zone traffic.
    • Receive traffic from local data plane proxies and local zone ingresses.
  • The zone ingress:
    • Receive XDS configuration from the local zone control plane.
    • Proxy traffic from other zone data plane proxies to local data plane proxies.
  • (optional) The zone egress:
    • Receive XDS configuration from the local zone control plane.
    • Proxy traffic from local data plane proxies:
      • to zone ingress proxies from other zones;
      • to external services from local zone;

Usage

To set up a multi-zone deployment we will need to:

Set up the global control plane

The global control plane must run on a dedicated cluster, and cannot be assigned to a zone.

The global control plane on Kubernetes must reside on its own Kubernetes cluster, to keep its resources separate from the resources the zone control planes create during synchronization.

  1. Run:

    1. kumactl install control-plane --mode=global | kubectl apply -f -
  2. Find the external IP and port of the global-remote-sync service in the kuma-system namespace:

    1. kubectl get services -n kuma-system
    2. NAMESPACE NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
    3. kuma-system global-remote-sync LoadBalancer 10.105.9.10 35.226.196.103 5685:30685/TCP 89s
    4. kuma-system kuma-control-plane ClusterIP 10.105.12.133 <none> 5681/TCP,443/TCP,5676/TCP,5677/TCP,5678/TCP,5679/TCP,5682/TCP,5653/UDP 90s

    In this example the value is 35.226.196.103:5685. You pass this as the value of <global-kds-address> when you set up the zone control planes.

  3. Set the controlPlane.mode value to global in the chart (values.yaml), then install. On the command line, run:

    1. helm install kuma --create-namespace --namespace kuma-system --set controlPlane.mode=global kuma/kuma

    Or you can edit the chart and pass the file to the helm install kuma command. To get the default values, run:

    1. helm show values kuma/kuma
  4. Find the external IP and port of the global-remote-sync service in the kuma-system namespace:

    1. kubectl get services -n kuma-system
    2. NAMESPACE NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
    3. kuma-system global-remote-sync LoadBalancer 10.105.9.10 35.226.196.103 5685:30685/TCP 89s
    4. kuma-system kuma-control-plane ClusterIP 10.105.12.133 <none> 5681/TCP,443/TCP,5676/TCP,5677/TCP,5678/TCP,5679/TCP,5682/TCP,5653/UDP 90s

    By default, it’s exposed on port 5685. In this example the value is 35.226.196.103:5685. You pass this as the value of <global-kds-address> when you set up the zone control planes.

  5. Set up the global control plane, and add the global environment variable:

    1. KUMA_MODE=global kuma-cp run

Set up the zone control planes

You need the following values to pass to each zone control plane setup:

  • zone – the zone name. An arbitrary string. This value registers the zone control plane with the global control plane.
  • kds-global-address – the external IP and port of the global control plane.

  • Kubernetes

  • Helm
  • Universal

Without zone egress:

  1. On each zone control plane, run:

    1. kumactl install control-plane \
    2. --mode=zone \
    3. --zone=<zone name> \
    4. --ingress-enabled \
    5. --kds-global-address grpcs://<global-kds-address>:5685 | kubectl apply -f -

    where zone is the same value for all zone control planes in the same zone.

With zone egress:

  1. On each zone control plane, run:

    1. kumactl install control-plane \
    2. --mode=zone \
    3. --zone=<zone-name> \
    4. --ingress-enabled \
    5. --egress-enabled \
    6. --kds-global-address grpcs://<global-kds-address>:5685 | kubectl apply -f -

    where zone is the same value for all zone control planes in the same zone.

Without zone egress:

  1. On each zone control plane, run:

    1. helm install kuma \
    2. --create-namespace \
    3. --namespace kuma-system \
    4. --set controlPlane.mode=zone \
    5. --set controlPlane.zone=<zone-name> \
    6. --set ingress.enabled=true \
    7. --set controlPlane.kdsGlobalAddress=grpcs://<global-kds-address>:5685 kuma/kuma

    where controlPlane.zone is the same value for all zone control planes in the same zone.

With zone egress:

  1. On each zone control plane, run:

    1. helm install kuma \
    2. --create-namespace \
    3. --namespace kuma-system \
    4. --set controlPlane.mode=zone \
    5. --set controlPlane.zone=<zone-name> \
    6. --set ingress.enabled=true \
    7. --set egress.enabled=true \
    8. --set controlPlane.kdsGlobalAddress=grpcs://<global-kds-address>:5685 kuma/kuma

    where controlPlane.zone is the same value for all zone control planes in the same zone.

  2. On each zone control plane, run:

    1. KUMA_MODE=zone \
    2. KUMA_MULTIZONE_ZONE_NAME=<zone-name> \
    3. KUMA_MULTIZONE_ZONE_GLOBAL_ADDRESS=grpcs://<global-kds-address>:5685 \
    4. ./kuma-cp run

    where KUMA_MULTIZONE_ZONE_NAME is the same value for all zone control planes in the same zone.

  3. Generate the zone ingress token:

    To register the zone ingress with the zone control plane, we need to generate a zone ingress token first

    1. kumactl generate zone-ingress-token --zone=<zone-name> > /tmp/ingress-token

    You can also generate the token with the REST API.

  4. Create an ingress data plane proxy configuration to allow kuma-cp services to be exposed for cross-zone communication:

    1. echo "type: ZoneIngress
    2. name: ingress-01
    3. networking:
    4. address: 127.0.0.1 # address that is routable within the zone
    5. port: 10000
    6. advertisedAddress: 10.0.0.1 # an address which other zones can use to consume this zone-ingress
    7. advertisedPort: 10000 # a port which other zones can use to consume this zone-ingress" > ingress-dp.yaml
  5. Apply the ingress config, passing the IP address of the zone control plane to cp-address:

    1. kuma-dp run \
    2. --proxy-type=ingress \
    3. --cp-address=https://<kuma-cp-address>:5678 \
    4. --dataplane-token-file=/tmp/ingress-token \
    5. --dataplane-file=ingress-dp.yaml

    If zone-ingress is running on a different machine than zone-cp you need to copy CA cert file from zone-cp (located in ~/.kuma/kuma-cp.crt) to somewhere accessible by zone-ingress (e.g. /tmp/kuma-cp.crt). Modify the above command and provide the certificate path in --ca-cert-file argument.

    1. kuma-dp run \
    2. --proxy-type=ingress \
    3. --cp-address=https://<kuma-cp-address>:5678 \
    4. --dataplane-token-file=/tmp/zone-token \
    5. --ca-cert-file=/tmp/kuma-cp.crt \
    6. --dataplane-file=ingress-dp.yaml
  6. Optional: if you want to deploy zone egress

    Create a ZoneEgress data plane proxy configuration to allow kuma-cp services

    To register the zone egress with the zone control plane, we need to generate a zone token first with appropriate scope first:

    1. kumactl generate zone-token --zone=<zone-name> --valid-for=24h --scope egress > /tmp/zoneegress-token

    You can also generate the token with the REST API.

  7. Create a ZoneEgress data plane proxy configuration to allow kuma-cp services to be configured to proxy traffic to other zones or external services through zone egress:

    1. echo "type: ZoneEgress
    2. name: zoneegress-01
    3. networking:
    4. address: 127.0.0.1 # address that is routable within the zone
    5. port: 10002" > zoneegress-dataplane.yaml
  8. Apply the egress config, passing the IP address of the zone control plane to cp-address:

    1. kuma-dp run \
    2. --proxy-type=egress \
    3. --cp-address=https://<kuma-cp-address>:5678 \
    4. --dataplane-token-file=/tmp/zoneegress-token \
    5. --dataplane-file=zoneegress-dataplane.yaml

Verify control plane connectivity

You can run kumactl get zones, or check the list of zones in the web UI for the global control plane, to verify zone control plane connections.

When a zone control plane connects to the global control plane, the Zone resource is created automatically in the global control plane.

The Zone Ingress tab of the web UI also lists zone control planes that you deployed with zone ingress.

Ensure mTLS is enabled on the multi-zone meshes

MTLS is mandatory to enable cross-zone service communication. mTLS can be configured in your mesh configuration as indicated in the mTLS section. This is required because Kuma uses the Server Name Indication field, part of the TLS protocol, as a way to pass routing information cross zones.

Cross-zone communication details

For this example we will assume we have a service running in a Kubernetes zone exposing a kuma.io/service with value echo-server_echo-example_svc_1010. The following examples are running in the remote zone trying to access the previously mentioned service.

To view the list of service names available, run:

  1. kubectl get serviceinsight all-services-default -oyaml
  2. apiVersion: kuma.io/v1alpha1
  3. kind: ServiceInsight
  4. mesh: default
  5. metadata:
  6. name: all-services-default
  7. spec:
  8. services:
  9. echo-server_echo-example_svc_1010:
  10. dataplanes:
  11. online: 1
  12. total: 1
  13. issuedBackends:
  14. ca-1: 1
  15. status: online

The following are some examples of different ways to address echo-server in the echo-example Namespace in a multi-zone mesh.

To send a request in the same zone, you can rely on Kubernetes DNS and use the usual Kubernetes hostnames and ports:

  1. curl http://echo-server:1010

Requests are distributed round robin between zones. You can use locality-aware load balancing to keep requests in the same zone.

To send a request to any zone, you can use the generated kuma.io/service and Kuma DNS:

  1. curl http://echo-server_echo-example_svc_1010.mesh:80

Kuma DNS also supports RFC 1123 compatible names, where underscores are replaced with dots:

  1. curl http://echo-server.echo-example.svc.1010.mesh:80
  1. kumactl inspect services
  2. SERVICE STATUS DATAPLANES
  3. echo-service_echo-example_svc_1010 Online 1/1

To consume the service in a Universal deployment without transparent proxy add the following outbound to your dataplane configuration:

  1. outbound:
  2. - port: 20012
  3. tags:
  4. kuma.io/service: echo-server_echo-example_svc_1010

From the data plane running you will now be able to reach the service using localhost:20012.

Alternatively, if you configure transparent proxy you can just call echo-server_echo-example_svc_1010.mesh without defining an outbound section.

For security reasons it’s not possible to customize the kuma.io/service in Kubernetes.

If you want to have the same service running on both Universal and Kubernetes make sure to align the Universal’s data plane inbound to have the same kuma.io/service as the one in Kubernetes or leverage TrafficRoute.

Failure modes

Global control plane offline

  • Policy updates will be impossible
  • Change in service list between zones will not propagate:
    • New services will not be discoverable in other zones.
    • Services removed from a zone will still appear available in other zones.
  • You won’t be able to disable or delete a zone.

Note that both local and cross-zone application traffic is not impacted by this failure case. Data plane proxy changes will be propagated within their zones.

Zone control plane offline

  • New data plane proxies won’t be able to join the mesh.
  • Data plane proxy configuration will not be updated.
  • Communication between data plane proxies will still work.
  • Cross zone communication will still work.
  • Other zones are unaffected.

You can think of this failure case as “Freezing” the zone mesh configuration. Communication will still work but changes will not be reflected on existing data plane proxies.

Communication between Global and Zone control plane failing

This can happen with misconfiguration or network connectivity issues between control planes.

  • Operations inside the zone will happen correctly (data plane proxies can join, leave and all configuration will be updated and sent correctly).
  • Policy changes will not be propagated to the zone control plane.
  • ZoneIngress, ZoneEgress and Dataplane changes will not be propagated to the global control plane:
    • The global inventory view of the data plane proxies will be outdated (this only impacts observability).
    • Other zones will not see new services registered inside this zone.
    • Other zones will not see services no longer running inside this zone.
    • Other zones will not see changes in number of instances of each service running in the local zone.
  • Global control plane will not send changes from other zone ingress to the zone:
    • Local data plane proxies will not see new services registered in other zones.
    • Local data plane proxies will not see services no longer running in other zones.
    • Local data plane proxies will not see changes in number of instances of each service running in other zones.
  • Global control plane will not send changes from other zone ingress to the zone.

Note that both local and cross-zone application traffic is not impacted by this failure case.

Communication between 2 zones failing

This can happen if there are network connectivity issues:

  • Between control plane and zone ingress from other zone.
  • Between control plane and zone egress (when present).
  • Between zone egress (when present) and zone ingress from other zone.
  • All Zone egress instances of a zone (when present) are down.
  • All zone ingress instances of a zone are down.

When it happens:

  • Communication and operation within each zone is unaffected.
  • Communication across each zone will fail.

With the right resiliency setup (Retries, Probes, Locality Aware LoadBalancing, Circuit Breakers) the failing zone can be quickly severed and traffic re-routed to another zone.

Delete a zone

To delete a Zone we must first shut down the corresponding Kuma zone control plane instances. As long as the Zone CP is running this will not be possible, and Kuma returns a validation error like:

  1. zone: unable to delete Zone, Zone CP is still connected, please shut it down first

When the Zone CP is fully disconnected and shut down, then the Zone can be deleted. All corresponding resources (like Dataplane and DataplaneInsight) will be deleted automatically as well.

  1. kubectl delete zone zone-1
  1. kumactl delete zone zone-1

Disable a zone

Change the enabled property value to false in the global control plane:

  1. apiVersion: kuma.io/v1alpha1
  2. kind: Zone
  3. metadata:
  4. name: zone-1
  5. spec:
  6. enabled: false
  1. type: Zone
  2. name: zone-1
  3. spec:
  4. enabled: false

With this setting, the global control plane will stop exchanging configuration with this zone. As a result, the zone’s ingress from zone-1 will be deleted from other zone and traffic won’t be routed to it anymore. The zone will show as Offline in the GUI and CLI.