Resource limits in fresh account do not match the default configuration #761
-
DescriptionThanks for your great starting point to get k8s up and running on hetzner cloud. However, module.kube-hetzner.null_resource.configure_autoscaler[0]: Creation complete after 1s [id=3475522426128264643]
╷
│ Error: server limit exceeded (resource_limit_exceeded)
│
│ with module.kube-hetzner.module.agents["2-0-storage"].hcloud_server.server,
│ on .terraform/modules/kube-hetzner/modules/host/main.tf line 22, in resource "hcloud_server" "server":
│ 22: resource "hcloud_server" "server" {
│
╵
╷
│ Error: server limit exceeded (resource_limit_exceeded)
│
│ with module.kube-hetzner.module.agents["3-0-egress"].hcloud_server.server,
│ on .terraform/modules/kube-hetzner/modules/host/main.tf line 22, in resource "hcloud_server" "server":
│ 22: resource "hcloud_server" "server" {
│
╵
╷
│ Error: server limit exceeded (resource_limit_exceeded)
│
│ with module.kube-hetzner.module.agents["1-0-agent-large"].hcloud_server.server,
│ on .terraform/modules/kube-hetzner/modules/host/main.tf line 22, in resource "hcloud_server" "server":
│ 22: resource "hcloud_server" "server" { My Account is too new to increase the limits - and I think a more leightweight setup would be sufficient. What would be a useful configuration to scale down resources to a minimum and still being able to test out Cilium/SPIFFE? Any hint to the right places in the configs would be appreciated :) My Kube.tf filecni_plugin = "cilium"
# ...
cilium_values = <<EOT
ipam:
mode: kubernetes
devices: "eth1"
k8s:
requireIPv4PodCIDR: true
kubeProxyReplacement: strict
l7Proxy: false
encryption:
enabled: true
type: wireguard
auth:
mTLS:
spire:
enabled: true
EOT ScreenshotsPlatformMac |
Beta Was this translation helpful? Give feedback.
Replies: 2 comments
-
Hey @manuschillerdev, just do not do HA, turn off both automatic upgrade (since HA is needed for that), go with 1 control plane, 1 agent, so that you have room for autoscalers. |
Beta Was this translation helpful? Give feedback.
Hey @manuschillerdev, just do not do HA, turn off both automatic upgrade (since HA is needed for that), go with 1 control plane, 1 agent, so that you have room for autoscalers.