feat(monitoring): enable Alertmanager (null receiver, longhorn PVC, az-a) + fix forgejo-rules ns forgejo->cicd — alerting delivery was disabled; forgejo PrometheusRule targeted a nonexistent namespace

This commit is contained in:
Story Crater Bot
2026-08-18 15:08:04 -07:00
parent 9e0a83f8ca
commit 61b906cce1
13 changed files with 145 additions and 139 deletions
+17 -8
View File
@@ -123,14 +123,23 @@ variable "controlplane_configs" {
variable "worker_configs" {
type = map(object({
hostname = string
lan_ip = string
lan_subnet = string
lan_gateway = string
install_disk = string
network_interface = optional(string, "eno1")
zone = string
gpu_count = optional(number, 0)
hostname = string
lan_ip = string
lan_subnet = string
lan_gateway = string
install_disk = string
network_interface = optional(string, "eno1")
zone = string
gpu_count = optional(number, 0)
# Extra node labels beyond the topology/GPU defaults.
node_labels = optional(map(string), {})
# Taints make a node dedicated: only pods carrying a matching toleration
# schedule there. effect is NoSchedule | PreferNoSchedule | NoExecute.
node_taints = optional(list(object({
key = string
value = string
effect = string
})), [])
factory_image = optional(string)
swap_size = optional(string, "")
ephemeral_max_size = optional(string, "700GiB")