Learn which pieces of networking infrastructure Pachyderm recommends setting up for production use.
March 22, 2023
To deploy in production, we recommend setting up the following pieces of networking infrastructure: A load balancer, a kubernetes ingress controller, and a DNS pointing to the load balancer. In addition we recommend using a managed database instance (such as RDS for AWS).
Interested in deploying with an embedded proxy and expose one single external port?
We are now shipping Pachyderm with an optional embedded proxy allowing your cluster to expose one single port externally. This deployment setup is optional.
If you choose to deploy Pachyderm with a Proxy, check out our new recommended architecture and deployment instructions.
Deploying with a proxy presents a couple of advantages:
- You only need to set up one TCP Load Balancer (No more Ingress in front of Console).
- You will need one DNS only.
- It simplifies the deployment of Console.
- No more port-forward.
Once you have your networking infrastructure set up, apply a helm values file such as the one specified in the example file below to wire up routing through an Ingress, and set up TLS. We recommend using a certificate manager such as cert-manager to refresh certificates and inject them as kubernetes secrets into your cluster for the ingress and load balancer to use.
This example uses Traefik as an Ingress controller. To configure other ingress controllers, apply their annotations in
values.yaml with enterprise license & authentication #
ingress: enabled: true host: <DNS-ENTRY> tls: enabled: true secretName: "pach-tls" pachd: tls: enabled: true secretName: "pach-tls" externalService: enabled: true loadBalancerIP: <LOAD-BALANCER-IP> console: enabled: true annotations: ## annotations specific to integrate with your ingress-controller ## the example below is a provided configuration specific to traefik as an ingress-controller traefik.ingress.kubernetes.io/router.tls: "true" kubernetes.io/ingress.class: "traefik"