How does Zeebe build a gateway cluster to achieve high availability. io. camunda. zeebe. client. ZeebeClientBuilder. Gatewayaddress (string) can only be configured with one gateway. How can a Zeebe hang without affecting its use. Can I provide deployment documents?

[bugbugmaker]

How does Zeebe build a gateway cluster to achieve high availability. io. camunda. zeebe. client. ZeebeClientBuilder. Gatewayaddress (string) can only be configured with one gateway. How can a Zeebe hang without affecting its use. Can I provide deployment documents?

[lenaschoenburg]

Hey @bugbugmaker 👋

If I understand correctly you have multiple gateways in you cluster and would like to configure the Java client to use more than one gateway for reliability in case gateways become unavailable.

You are right that this is currently not supported in the Java client (or any other client as far as I know). We generally recommend to put multiple gateways behind some sort of load balancing. This hides all the complexity of handling multiple gateways from the client and also allows for dynamically changing the number of gateways as needed without adjusting the client configuration. This is also exactly what we do on Camunda Cloud.

Would this be an option for you?

[bugbugmaker]

Hey @bugbugmaker 👋

If I understand correctly you have multiple gateways in you cluster and would like to configure the Java client to use more than one gateway for reliability in case gateways become unavailable.

You are right that this is currently not supported in the Java client (or any other client as far as I know). We generally recommend to put multiple gateways behind some sort of load balancing. This hides all the complexity of handling multiple gateways from the client and also allows for dynamically changing the number of gateways as needed without adjusting the client configuration. This is also exactly what we do on Camunda Cloud.

Would this be an option for you?

"We generally recommend to put multiple gateways behind some sort of load balancing. "
I don't quite understand what this means. Can you explain it in detail?

[lenaschoenburg]

Let's take the official helm charts for example. There, all gateways are accessible from the same Kubernetes service. The client is then configured for the k8s service and doesn't need to know about individual gateways.

We also have two places that are a bit better suited for questions: The community forum and slack community

[Zelldon]

I think the question is answered so I will close this for now. If something is unclear you can always reopen it.