Are there any issues with running multiple replicas of Keto for High availability on K8s? Can I just edit the helm values to bringup 2 replicas?

Yep, it'll work fine, you just need to set

replicaCount

in you helm values, or enable autoscaling with something like the below

deployment:
  autoscaling:
    enabled: true
    minReplicas: 3

Awesome. Thanks. Are there special considerations for affinity rules over replica count?

They should be used independently rather either or. The affinity rules would be to just ensure you don't schedule all your Keto pods on a single node, or zone, giving better high availability

Thanks @tall-angle-41306

Which cloud provider are you running on?

Using EKS from AWS

This should work for EKS:

affinity:
  podAntiAffinity:
    requiredDuringSchedulingIgnoredDuringExecution:
    - labelSelector:
        matchExpressions:
        - key: <http://app.kubernetes.io/instance|app.kubernetes.io/instance>
          operator: In
          values:
          - keto
      topologyKey: <http://kubernetes.io/hostname|kubernetes.io/hostname>
    preferredDuringSchedulingIgnoredDuringExecution:
    - weight: 100
      podAffinityTerm:
        labelSelector:
        - key: <http://app.kubernetes.io/instance|app.kubernetes.io/instance>
          operator: In
          values:
          - keto
        topologyKey: <http://topology.kubernetes.io/zone|topology.kubernetes.io/zone>

The above will ensure your Keto pods don't share the same node, and will prefer (but not require) to be split across any of your zones

Thanks @tall-angle-41306 Will try this out