infra/stacks/kyverno/modules/kyverno/main.tf
Viktor Barzin 0216e993dc
Some checks failed
ci/woodpecker/push/build-cli Pipeline was successful
ci/woodpecker/push/default Pipeline failed
etcd-load-reduction: remove VPA/Goldilocks, disable kyverno reporting, descheduler hourly
The control-plane flap (etcd lease-renewal timeouts) recurred. Rather than move
etcd to SSD (code-oflt, deferred again), the chosen direction is to REDUCE etcd
load enough that the leader-election-timeout band-aid (renew 10s->30s) becomes
removable. These are the big, clean cuts:

1. Remove VPA/Goldilocks (stacks/vpa emptied). All 349 VPAs ran updateMode=Off
   (no auto-right-sizing) yet cost ~800 etcd objects + continuous recommender
   writes + a pod-creation admission webhook, purely to feed a dashboard. krr
   (Dockerized, on-demand) replaces it. Reverses the re-add after memory 2431.

2. Disable kyverno reporting (admission/aggregate/background). policyReports were
   already off, so the pipeline generated ephemeralreports + an hourly
   all-resource etcd re-scan for NO user-facing output. Admission enforcement
   (deny-* policies) and Keel mutation are unaffected; violations surface via
   Loki->Slack.

3. descheduler */5 -> hourly (fewer list/evict cycles; rebalancing isn't urgent).

Deferred (poor ROI / unsafe as planned): ESO refreshInterval 15m->1h is a
~20-stack sprawl for ~0.1 writes/s; keel background=false is invalid for a
mutate-existing policy and its churn is apply-time not steady-state. Both filed
as follow-up beads.

Post-apply: delete the chart-orphaned VPA CRDs to cascade-clean leftover CRs.
Then measure etcd apply-latency and revert the timeouts. Docs updated
(VPA/Goldilocks -> krr). See memory 5402-5407.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-12 19:41:22 +00:00

241 lines
6.2 KiB
HCL

resource "kubernetes_namespace" "kyverno" {
metadata {
name = "kyverno"
labels = {
"istio-injection" : "disabled"
"keel.sh/enrolled" = "true"
}
}
lifecycle {
# KYVERNO_LIFECYCLE_V1: goldilocks-vpa-auto-mode ClusterPolicy stamps this label on every namespace
ignore_changes = [metadata[0].labels["goldilocks.fairwinds.com/vpa-update-mode"]]
}
}
resource "helm_release" "kyverno" {
namespace = kubernetes_namespace.kyverno.metadata[0].name
create_namespace = false
name = "kyverno"
atomic = true
repository = "https://kyverno.github.io/kyverno/"
chart = "kyverno"
version = "3.6.1"
values = [yamlencode({
# When Kyverno is unavailable, allow pod creation to proceed without
# mutation/validation rather than blocking all admissions cluster-wide.
features = {
forceFailurePolicyIgnore = {
enabled = true
}
# Reporting fully disabled (2026-06-12, etcd-load-reduction). policyReports
# were already off, so admission/aggregate/background reporting generated
# ephemeralreports + an hourly all-resource etcd re-scan for NO user-facing
# output. Admission enforcement (deny-* policies) and Keel mutation are
# independent of reporting; policy violations surface via Loki->Slack. This
# removes a steady-state etcd write/scan load (control-plane flap mitigation).
policyReports = {
enabled = false
}
admissionReports = {
enabled = false
}
aggregateReports = {
enabled = false
}
backgroundScan = {
enabled = false
}
}
reportsController = {
resources = {
limits = {
memory = "512Mi"
}
requests = {
cpu = "100m"
memory = "384Mi"
}
}
}
backgroundController = {
resources = {
# Bumped 2026-05-16 from 384Mi → 2Gi because the controller OOMKilled
# while processing 176 UpdateRequests for the inject-keel-annotations
# mutate-existing scan. With mutateExistingOnPolicyUpdate=true the
# background controller needs significantly more memory during the
# initial bulk scan.
limits = {
memory = "2Gi"
}
requests = {
cpu = "100m"
memory = "256Mi"
}
}
}
cleanupController = {
resources = {
limits = {
memory = "192Mi"
}
requests = {
cpu = "100m"
memory = "192Mi"
}
}
}
admissionController = {
replicas = 2
updateStrategy = {
type = "RollingUpdate"
rollingUpdate = {
maxSurge = 0
maxUnavailable = 1
}
}
container = {
resources = {
limits = {
memory = "256Mi"
}
requests = {
cpu = "100m"
memory = "256Mi"
}
}
}
# More tolerant liveness probe — API server slowness shouldn't kill the pod
livenessProbe = {
httpGet = {
path = "/health/liveness"
port = 9443
scheme = "HTTPS"
}
initialDelaySeconds = 15
periodSeconds = 30
timeoutSeconds = 5
failureThreshold = 4
successThreshold = 1
}
# Spread replicas across nodes for HA
topologySpreadConstraints = [
{
maxSkew = 1
topologyKey = "kubernetes.io/hostname"
whenUnsatisfiable = "DoNotSchedule"
labelSelector = {
matchLabels = {
"app.kubernetes.io/component" = "admission-controller"
"app.kubernetes.io/instance" = "kyverno"
}
}
}
]
}
})]
}
# To unlabel all:
# kubectl label deployment,statefulset,daemonset --all-namespaces -l tier tier-
#
# Uses namespaceSelector to match tiers — no API call needed.
# One rule per tier so Kyverno resolves the tier value from its informer cache.
resource "kubectl_manifest" "mutate_tier_from_namespace" {
yaml_body = yamlencode({
apiVersion = "kyverno.io/v1"
kind = "ClusterPolicy"
metadata = {
name = "sync-tier-label-from-namespace"
}
spec = {
rules = [for tier in local.governance_tiers : {
name = "sync-tier-${tier}"
match = {
any = [
{
resources = {
kinds = ["Deployment", "StatefulSet", "DaemonSet"]
namespaceSelector = {
matchLabels = {
tier = tier
}
}
}
}
]
}
exclude = {
any = [
{
resources = {
namespaces = ["kube-system", "metallb-system", "n8n"]
}
}
]
}
mutate = {
patchStrategicMerge = {
metadata = {
labels = {
"+(tier)" = tier
}
}
}
}
}]
}
})
}
# resource "kubectl_manifest" "enforce_pod_tier_label" {
# manifest = {
# apiVersion = "kyverno.io/v1"
# kind = "ClusterPolicy"
# metadata = {
# name = "enforce-pod-tier-label"
# annotations = {
# "policies.kyverno.io/description" = "Rejects any pod that does not have a tier label."
# }
# }
# spec = {
# # 'Enforce' blocks the creation. 'Audit' just reports it.
# validationFailureAction = "Enforce"
# background = true
# rules = [
# {
# name = "check-for-tier-label"
# match = {
# any = [
# {
# resources = {
# kinds = ["Pod"]
# }
# }
# ]
# }
# validate = {
# message = "The label 'tier' is required for all pods in this cluster."
# pattern = {
# metadata = {
# labels = {
# "tier" = "?*" # The "?*" syntax means the value must not be empty
# }
# }
# }
# }
# }
# ]
# }
# }
# }