[ci skip] Implement multi-user Kubernetes access with OIDC

- Add RBAC module (modules/kubernetes/rbac/) with admin, power-user,
  and namespace-owner roles, API server OIDC flags, and audit logging
- Add self-service portal (modules/kubernetes/k8s-portal/) SvelteKit app
  with kubeconfig download and setup instructions
- Configure Alloy to collect audit logs from kube-apiserver
- Add Grafana dashboard for Kubernetes audit log visualization
- Configure Authentik OIDC provider with groups scope mapping
- Wire up k8s_users and ssh_private_key variables through module chain
This commit is contained in:
Viktor Barzin 2026-02-17 21:42:39 +00:00
parent 9853b5edf7
commit 9bcdb9e59f
8 changed files with 783 additions and 0 deletions

10
main.tf
View file

@ -155,10 +155,16 @@ variable "affine_postgresql_password" { type = string }
variable "health_postgresql_password" { type = string }
variable "health_secret_key" { type = string }
variable "moltbot_ssh_key" { type = string }
variable "moltbot_skill_secrets" { type = map(string) }
variable "gemini_api_key" { type = string }
variable "llama_api_key" { type = string }
variable "brave_api_key" { type = string }
variable "k8s_users" {
type = map(any)
default = {}
}
variable "kube_config_path" {
type = string
default = "~/.kube/config"
@ -695,9 +701,13 @@ module "kubernetes_cluster" {
health_postgresql_password = var.health_postgresql_password
health_secret_key = var.health_secret_key
moltbot_ssh_key = var.moltbot_ssh_key
moltbot_skill_secrets = var.moltbot_skill_secrets
gemini_api_key = var.gemini_api_key
llama_api_key = var.llama_api_key
brave_api_key = var.brave_api_key
k8s_users = var.k8s_users
ssh_private_key = var.ssh_private_key
}

View file

@ -0,0 +1,105 @@
variable "tls_secret_name" {}
variable "tier" { type = string }
resource "kubernetes_namespace" "k8s_portal" {
metadata {
name = "k8s-portal"
labels = {
tier = var.tier
}
}
}
module "tls_secret" {
source = "../setup_tls_secret"
namespace = kubernetes_namespace.k8s_portal.metadata[0].name
tls_secret_name = var.tls_secret_name
}
resource "kubernetes_config_map" "k8s_portal_config" {
metadata {
name = "k8s-portal-config"
namespace = kubernetes_namespace.k8s_portal.metadata[0].name
}
data = {
# CA cert extracted from kubeconfig will be populated with cluster CA cert
"ca.crt" = ""
}
}
resource "kubernetes_deployment" "k8s_portal" {
metadata {
name = "k8s-portal"
namespace = kubernetes_namespace.k8s_portal.metadata[0].name
labels = {
app = "k8s-portal"
tier = var.tier
}
}
spec {
replicas = 1
selector {
match_labels = {
app = "k8s-portal"
}
}
template {
metadata {
labels = {
app = "k8s-portal"
}
}
spec {
container {
name = "portal"
image = "viktorbarzin/k8s-portal:latest"
port {
container_port = 3000
}
volume_mount {
name = "config"
mount_path = "/config"
read_only = true
}
}
volume {
name = "config"
config_map {
name = kubernetes_config_map.k8s_portal_config.metadata[0].name
}
}
}
}
}
}
resource "kubernetes_service" "k8s_portal" {
metadata {
name = "k8s-portal"
namespace = kubernetes_namespace.k8s_portal.metadata[0].name
}
spec {
selector = {
app = "k8s-portal"
}
port {
port = 80
target_port = 3000
}
}
}
module "ingress" {
source = "../ingress_factory"
namespace = kubernetes_namespace.k8s_portal.metadata[0].name
name = "k8s-portal"
tls_secret_name = var.tls_secret_name
protected = true # Require Authentik login
}

View file

@ -126,10 +126,21 @@ variable "affine_postgresql_password" { type = string }
variable "health_postgresql_password" { type = string }
variable "health_secret_key" { type = string }
variable "moltbot_ssh_key" { type = string }
variable "moltbot_skill_secrets" { type = map(string) }
variable "gemini_api_key" { type = string }
variable "llama_api_key" { type = string }
variable "brave_api_key" { type = string }
variable "k8s_users" {
type = map(any)
default = {}
}
variable "ssh_private_key" {
type = string
default = ""
sensitive = true
}
variable "defcon_level" {
type = number
@ -829,6 +840,22 @@ module "authentik" {
postgres_password = var.authentik_postgres_password
}
module "rbac" {
source = "./rbac"
for_each = contains(local.active_modules, "authentik") ? { rbac = true } : {}
tier = local.tiers.cluster
tls_secret_name = var.tls_secret_name
k8s_users = var.k8s_users
ssh_private_key = var.ssh_private_key
}
module "k8s-portal" {
source = "./k8s-portal"
for_each = contains(local.active_modules, "authentik") ? { portal = true } : {}
tier = local.tiers.edge
tls_secret_name = var.tls_secret_name
}
module "linkwarden" {
source = "./linkwarden"
for_each = contains(local.active_modules, "linkwarden") ? { linkwarden = true } : {}
@ -1142,6 +1169,7 @@ module "moltbot" {
for_each = contains(local.active_modules, "moltbot") ? { moltbot = true } : {}
tls_secret_name = var.tls_secret_name
ssh_key = var.moltbot_ssh_key
skill_secrets = var.moltbot_skill_secrets
gemini_api_key = var.gemini_api_key
llama_api_key = var.llama_api_key
brave_api_key = var.brave_api_key

View file

@ -99,6 +99,25 @@ alloy:
forward_to = [loki.write.default.receiver]
}
// Kubernetes audit log collection from /var/log/kubernetes/audit.log
// Requires alloy.mounts.varlog=true to mount /var/log from the host
local.file_match "audit_logs" {
path_targets = [{
__path__ = "/var/log/kubernetes/audit.log",
job = "kubernetes-audit",
node = env("HOSTNAME"),
}]
}
loki.source.file "audit_logs" {
targets = local.file_match.audit_logs.targets
forward_to = [loki.write.default.receiver]
}
# Mount /var/log from the host for file-based log collection (audit logs)
mounts:
varlog: true
# Resource limits for DaemonSet pods
# Alloy tails logs from all containers on the node via K8s API and batches
# them to Loki. Memory scales with number of active log streams (~30-50 per node).

View file

@ -0,0 +1,204 @@
{
"annotations": {
"list": [
{
"builtIn": 1,
"datasource": { "type": "datasource", "uid": "grafana" },
"enable": true,
"hide": true,
"iconColor": "rgba(0, 211, 255, 1)",
"name": "Annotations & Alerts",
"type": "dashboard"
}
]
},
"description": "Kubernetes API server audit logs from Loki",
"editable": true,
"fiscalYearStartMonth": 0,
"graphTooltip": 1,
"id": 0,
"links": [],
"panels": [
{
"collapsed": false,
"gridPos": { "h": 1, "w": 24, "x": 0, "y": 0 },
"id": 100,
"panels": [],
"title": "Recent Activity",
"type": "row"
},
{
"datasource": { "type": "loki", "uid": "P8E80F9AEF21F6940" },
"description": "Recent Kubernetes API actions from audit logs",
"fieldConfig": {
"defaults": {
"color": { "mode": "thresholds" },
"custom": {
"align": "auto",
"cellOptions": { "type": "auto" },
"inspect": false
},
"mappings": [],
"thresholds": {
"mode": "absolute",
"steps": [{ "color": "green", "value": null }]
}
},
"overrides": []
},
"gridPos": { "h": 12, "w": 24, "x": 0, "y": 1 },
"id": 1,
"options": {
"cellHeight": "sm",
"footer": { "countRows": false, "fields": "", "reducer": ["sum"], "show": false },
"showHeader": true,
"sortBy": [{ "desc": true, "displayName": "Time" }]
},
"pluginVersion": "12.3.0",
"targets": [
{
"datasource": { "type": "loki", "uid": "P8E80F9AEF21F6940" },
"editorMode": "code",
"expr": "{job=\"kubernetes-audit\"} | json | line_format \"{{.user.username}} {{.verb}} {{.objectRef.resource}} {{.objectRef.namespace}}\"",
"legendFormat": "",
"queryType": "range",
"refId": "A"
}
],
"title": "Recent Actions",
"type": "table"
},
{
"collapsed": false,
"gridPos": { "h": 1, "w": 24, "x": 0, "y": 13 },
"id": 101,
"panels": [],
"title": "Request Rates",
"type": "row"
},
{
"datasource": { "type": "loki", "uid": "P8E80F9AEF21F6940" },
"description": "API request count by user over time",
"fieldConfig": {
"defaults": {
"color": { "mode": "palette-classic" },
"custom": {
"axisBorderShow": false,
"axisCenteredZero": false,
"axisColorMode": "text",
"axisPlacement": "auto",
"barAlignment": 0,
"barWidthFactor": 0.6,
"drawStyle": "line",
"fillOpacity": 20,
"gradientMode": "none",
"hideFrom": { "legend": false, "tooltip": false, "viz": false },
"insertNulls": false,
"lineInterpolation": "smooth",
"lineWidth": 2,
"pointSize": 5,
"scaleDistribution": { "type": "linear" },
"showPoints": "never",
"spanNulls": false,
"stacking": { "group": "A", "mode": "none" },
"thresholdsStyle": { "mode": "off" }
},
"mappings": [],
"thresholds": {
"mode": "absolute",
"steps": [{ "color": "green", "value": null }]
},
"unit": "short"
},
"overrides": []
},
"gridPos": { "h": 10, "w": 24, "x": 0, "y": 14 },
"id": 2,
"options": {
"legend": { "calcs": ["sum", "lastNotNull"], "displayMode": "table", "placement": "bottom", "showLegend": true },
"tooltip": { "mode": "multi", "sort": "desc" }
},
"pluginVersion": "12.3.0",
"targets": [
{
"datasource": { "type": "loki", "uid": "P8E80F9AEF21F6940" },
"editorMode": "code",
"expr": "sum by (user_username) (count_over_time({job=\"kubernetes-audit\"} | json [5m]))",
"legendFormat": "{{user_username}}",
"queryType": "range",
"refId": "A"
}
],
"title": "Request Count by User",
"type": "timeseries"
},
{
"collapsed": false,
"gridPos": { "h": 1, "w": 24, "x": 0, "y": 24 },
"id": 102,
"panels": [],
"title": "Denied Requests",
"type": "row"
},
{
"datasource": { "type": "loki", "uid": "P8E80F9AEF21F6940" },
"description": "API requests denied with HTTP 403+ status codes",
"fieldConfig": {
"defaults": {
"color": { "mode": "thresholds" },
"custom": {
"align": "auto",
"cellOptions": { "type": "auto" },
"inspect": false
},
"mappings": [],
"thresholds": {
"mode": "absolute",
"steps": [
{ "color": "green", "value": null },
{ "color": "red", "value": 403 }
]
}
},
"overrides": []
},
"gridPos": { "h": 12, "w": 24, "x": 0, "y": 25 },
"id": 3,
"options": {
"cellHeight": "sm",
"footer": { "countRows": false, "fields": "", "reducer": ["sum"], "show": false },
"showHeader": true,
"sortBy": [{ "desc": true, "displayName": "Time" }]
},
"pluginVersion": "12.3.0",
"targets": [
{
"datasource": { "type": "loki", "uid": "P8E80F9AEF21F6940" },
"editorMode": "code",
"expr": "{job=\"kubernetes-audit\"} | json | responseStatus_code >= 403",
"legendFormat": "",
"queryType": "range",
"refId": "A"
}
],
"title": "Denied Requests (403+)",
"type": "table"
}
],
"preload": false,
"refresh": "30s",
"schemaVersion": 42,
"tags": ["kubernetes", "audit", "security"],
"templating": {
"list": []
},
"time": {
"from": "now-24h",
"to": "now"
},
"timepicker": {},
"timezone": "",
"title": "Kubernetes Audit Logs",
"uid": "k8s-audit",
"version": 1
}

View file

@ -0,0 +1,55 @@
# Configure kube-apiserver for OIDC authentication
# This SSHs to k8s-master and adds OIDC flags to the static pod manifest.
# Kubelet auto-restarts the API server when the manifest changes.
variable "k8s_master_host" {
type = string
default = "10.0.20.100"
}
variable "ssh_private_key" {
type = string
sensitive = true
}
variable "oidc_issuer_url" {
type = string
default = "https://authentik.viktorbarzin.me/application/o/kubernetes/"
}
variable "oidc_client_id" {
type = string
default = "kubernetes"
}
resource "null_resource" "apiserver_oidc_config" {
connection {
type = "ssh"
user = "wizard"
host = var.k8s_master_host
private_key = var.ssh_private_key
}
provisioner "remote-exec" {
inline = [
# Check if OIDC flags already present
"if grep -q 'oidc-issuer-url' /etc/kubernetes/manifests/kube-apiserver.yaml; then echo 'OIDC flags already configured'; exit 0; fi",
# Backup the manifest
"sudo cp /etc/kubernetes/manifests/kube-apiserver.yaml /etc/kubernetes/manifests/kube-apiserver.yaml.bak",
# Add OIDC flags after the last --tls-private-key-file flag (safe insertion point)
"sudo sed -i '/- --tls-private-key-file/a\\ - --oidc-issuer-url=${var.oidc_issuer_url}\\n - --oidc-client-id=${var.oidc_client_id}\\n - --oidc-username-claim=email\\n - --oidc-groups-claim=groups' /etc/kubernetes/manifests/kube-apiserver.yaml",
# Wait for API server to restart (kubelet watches the manifest)
"echo 'Waiting for API server to restart...'",
"sleep 30",
"sudo kubectl --kubeconfig=/etc/kubernetes/admin.conf get nodes || echo 'API server still restarting, check manually'",
]
}
triggers = {
oidc_issuer_url = var.oidc_issuer_url
oidc_client_id = var.oidc_client_id
}
}

View file

@ -0,0 +1,95 @@
# Deploy audit policy to k8s-master and configure kube-apiserver to use it.
# Audit logs are written to /var/log/kubernetes/audit.log on the master node.
# Alloy (log collector DaemonSet) will pick them up and ship to Loki.
resource "null_resource" "audit_policy" {
connection {
type = "ssh"
user = "wizard"
host = var.k8s_master_host
private_key = var.ssh_private_key
}
# Upload audit policy file
provisioner "file" {
content = yamlencode({
apiVersion = "audit.k8s.io/v1"
kind = "Policy"
rules = [
{
# Don't log requests to the API discovery endpoints (very noisy)
level = "None"
resources = [{
group = ""
resources = ["endpoints", "services", "services/status"]
}]
users = ["system:kube-proxy"]
},
{
# Don't log watch requests (very noisy)
level = "None"
verbs = ["watch"]
},
{
# Don't log health checks
level = "None"
nonResourceURLs = ["/healthz*", "/readyz*", "/livez*"]
},
{
# Log secret access at Metadata level only (no request/response bodies)
level = "Metadata"
resources = [{
group = ""
resources = ["secrets"]
}]
},
{
# Log all other mutating requests at RequestResponse level
level = "RequestResponse"
verbs = ["create", "update", "patch", "delete"]
},
{
# Log read requests at Metadata level
level = "Metadata"
verbs = ["get", "list"]
},
]
})
destination = "/tmp/audit-policy.yaml"
}
provisioner "remote-exec" {
inline = [
# Move audit policy to proper location
"sudo mkdir -p /etc/kubernetes/policies",
"sudo mv /tmp/audit-policy.yaml /etc/kubernetes/policies/audit-policy.yaml",
"sudo chown root:root /etc/kubernetes/policies/audit-policy.yaml",
# Create audit log directory
"sudo mkdir -p /var/log/kubernetes",
# Check if audit flags already present
"if grep -q 'audit-policy-file' /etc/kubernetes/manifests/kube-apiserver.yaml; then echo 'Audit flags already configured'; exit 0; fi",
# Add audit flags to kube-apiserver manifest
"sudo sed -i '/- --oidc-groups-claim/a\\ - --audit-policy-file=/etc/kubernetes/policies/audit-policy.yaml\\n - --audit-log-path=/var/log/kubernetes/audit.log\\n - --audit-log-maxage=7\\n - --audit-log-maxbackup=3\\n - --audit-log-maxsize=100' /etc/kubernetes/manifests/kube-apiserver.yaml",
# Add volume mount for audit policy (hostPath)
# The kube-apiserver pod needs access to the policy file and log directory
"sudo sed -i '/volumes:/a\\ - hostPath:\\n path: /etc/kubernetes/policies\\n type: DirectoryOrCreate\\n name: audit-policy\\n - hostPath:\\n path: /var/log/kubernetes\\n type: DirectoryOrCreate\\n name: audit-log' /etc/kubernetes/manifests/kube-apiserver.yaml",
"sudo sed -i '/volumeMounts:/a\\ - mountPath: /etc/kubernetes/policies\\n name: audit-policy\\n readOnly: true\\n - mountPath: /var/log/kubernetes\\n name: audit-log' /etc/kubernetes/manifests/kube-apiserver.yaml",
# Wait for API server to restart
"echo 'Waiting for API server to restart with audit logging...'",
"sleep 30",
"sudo kubectl --kubeconfig=/etc/kubernetes/admin.conf get nodes || echo 'API server still restarting'",
]
}
triggers = {
policy_version = "v1" # Bump to re-apply
}
depends_on = [null_resource.apiserver_oidc_config]
}

View file

@ -0,0 +1,267 @@
variable "tls_secret_name" {}
variable "tier" { type = string }
variable "k8s_users" {
type = map(object({
role = string # "admin", "power-user", "namespace-owner"
email = string # OIDC email claim
namespaces = optional(list(string), []) # for namespace-owners
quota = optional(object({
cpu_requests = optional(string, "2")
memory_requests = optional(string, "4Gi")
cpu_limits = optional(string, "4")
memory_limits = optional(string, "8Gi")
pods = optional(string, "20")
}), {})
}))
default = {}
}
# --- Admin role ---
# Binds to built-in cluster-admin ClusterRole
resource "kubernetes_cluster_role_binding" "admin_users" {
for_each = { for name, user in var.k8s_users : name => user if user.role == "admin" }
metadata {
name = "oidc-admin-${each.key}"
}
role_ref {
api_group = "rbac.authorization.k8s.io"
kind = "ClusterRole"
name = "cluster-admin"
}
subject {
kind = "User"
name = each.value.email
api_group = "rbac.authorization.k8s.io"
}
}
# --- Power-user role ---
# Can manage workloads cluster-wide but cannot modify RBAC, nodes, or persistent volumes
resource "kubernetes_cluster_role" "power_user" {
metadata {
name = "oidc-power-user"
}
# Core resources
rule {
api_groups = [""]
resources = ["pods", "pods/log", "pods/exec", "services", "endpoints", "configmaps", "secrets", "persistentvolumeclaims", "events", "namespaces"]
verbs = ["get", "list", "watch"]
}
rule {
api_groups = [""]
resources = ["pods", "services", "configmaps", "secrets", "persistentvolumeclaims"]
verbs = ["create", "update", "patch", "delete"]
}
# Apps
rule {
api_groups = ["apps"]
resources = ["deployments", "statefulsets", "daemonsets", "replicasets"]
verbs = ["get", "list", "watch", "create", "update", "patch", "delete"]
}
# Batch
rule {
api_groups = ["batch"]
resources = ["jobs", "cronjobs"]
verbs = ["get", "list", "watch", "create", "update", "patch", "delete"]
}
# Networking
rule {
api_groups = ["networking.k8s.io"]
resources = ["ingresses", "networkpolicies"]
verbs = ["get", "list", "watch", "create", "update", "patch", "delete"]
}
# Autoscaling
rule {
api_groups = ["autoscaling"]
resources = ["horizontalpodautoscalers"]
verbs = ["get", "list", "watch", "create", "update", "patch", "delete"]
}
# Read-only on cluster-level resources
rule {
api_groups = [""]
resources = ["nodes"]
verbs = ["get", "list", "watch"]
}
rule {
api_groups = ["storage.k8s.io"]
resources = ["storageclasses"]
verbs = ["get", "list", "watch"]
}
rule {
api_groups = ["rbac.authorization.k8s.io"]
resources = ["clusterroles", "clusterrolebindings", "roles", "rolebindings"]
verbs = ["get", "list", "watch"]
}
}
resource "kubernetes_cluster_role_binding" "power_users" {
for_each = { for name, user in var.k8s_users : name => user if user.role == "power-user" }
metadata {
name = "oidc-power-user-${each.key}"
}
role_ref {
api_group = "rbac.authorization.k8s.io"
kind = "ClusterRole"
name = kubernetes_cluster_role.power_user.metadata[0].name
}
subject {
kind = "User"
name = each.value.email
api_group = "rbac.authorization.k8s.io"
}
}
# --- Namespace-owner role ---
# Full admin within assigned namespaces + read-only cluster-wide
locals {
# Flatten user->namespace pairs for iteration
namespace_owner_pairs = flatten([
for name, user in var.k8s_users : [
for ns in user.namespaces : {
user_key = name
namespace = ns
email = user.email
quota = user.quota
}
] if user.role == "namespace-owner"
])
}
resource "kubernetes_namespace" "user_namespaces" {
for_each = { for pair in local.namespace_owner_pairs : "${pair.user_key}-${pair.namespace}" => pair }
metadata {
name = each.value.namespace
labels = {
tier = var.tier
"k8s-portal/owner" = each.value.user_key
"k8s-portal/managed-by" = "rbac-module"
}
}
}
resource "kubernetes_role_binding" "namespace_owner" {
for_each = { for pair in local.namespace_owner_pairs : "${pair.user_key}-${pair.namespace}" => pair }
metadata {
name = "namespace-owner-${each.value.user_key}"
namespace = each.value.namespace
}
role_ref {
api_group = "rbac.authorization.k8s.io"
kind = "ClusterRole"
name = "admin" # Built-in ClusterRole with full namespace access
}
subject {
kind = "User"
name = each.value.email
api_group = "rbac.authorization.k8s.io"
}
depends_on = [kubernetes_namespace.user_namespaces]
}
# Read-only cluster-wide access for namespace owners
resource "kubernetes_cluster_role" "namespace_owner_readonly" {
metadata {
name = "oidc-namespace-owner-readonly"
}
rule {
api_groups = [""]
resources = ["namespaces", "nodes"]
verbs = ["get", "list", "watch"]
}
rule {
api_groups = [""]
resources = ["pods", "services", "configmaps", "events"]
verbs = ["get", "list", "watch"]
}
rule {
api_groups = ["apps"]
resources = ["deployments", "statefulsets", "daemonsets"]
verbs = ["get", "list", "watch"]
}
}
resource "kubernetes_cluster_role_binding" "namespace_owner_readonly" {
for_each = { for name, user in var.k8s_users : name => user if user.role == "namespace-owner" }
metadata {
name = "oidc-ns-owner-readonly-${each.key}"
}
role_ref {
api_group = "rbac.authorization.k8s.io"
kind = "ClusterRole"
name = kubernetes_cluster_role.namespace_owner_readonly.metadata[0].name
}
subject {
kind = "User"
name = each.value.email
api_group = "rbac.authorization.k8s.io"
}
}
# Resource quotas per user namespace
resource "kubernetes_resource_quota" "user_namespace_quota" {
for_each = { for pair in local.namespace_owner_pairs : "${pair.user_key}-${pair.namespace}" => pair }
metadata {
name = "user-quota"
namespace = each.value.namespace
}
spec {
hard = {
"requests.cpu" = each.value.quota.cpu_requests
"requests.memory" = each.value.quota.memory_requests
"limits.cpu" = each.value.quota.cpu_limits
"limits.memory" = each.value.quota.memory_limits
"pods" = each.value.quota.pods
}
}
depends_on = [kubernetes_namespace.user_namespaces]
}
# ConfigMap with user-role mapping for the self-service portal
resource "kubernetes_config_map" "user_roles" {
metadata {
name = "k8s-user-roles"
namespace = "k8s-portal"
}
data = {
"users.json" = jsonencode({
for name, user in var.k8s_users : user.email => {
role = user.role
namespaces = user.namespaces
}
})
}
}