Files
Scooby HuskyandClaude Sonnet 5 7a6ecd14fe Bump rancher memory limit 2Gi -> 3Gi, was OOMKilling during restart recovery
Repeated restarts from tonight's network disruption kept OOMKilling
rancher - each restart's startup reconciliation is memory-hungry, and 2Gi
wasn't enough headroom for that burst even though steady-state usage is
normally fine. Nodes have ample free memory (26-36% used).

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
2026-08-17 21:01:48 -05:00

53 lines
1.5 KiB
YAML

# apps/rancher/values.yaml
# Rancher - Kubernetes Management Platform with Authentik SSO
hostname: rancher.kube.huskypup.net
# Exposed via Istio edge Gateway + VirtualService (apps/rancher/istio-virtualservice.yaml)
ingress:
enabled: false
# TLS terminates at Istio ingressgateway
tls: external
# Single replica for homelab (scale up if needed)
replicas: 1
# Storage class for Rancher's internal state
# Note: Rancher itself is mostly stateless; state is in etcd/k8s
resources:
requests:
cpu: 50m
memory: 512Mi
limits:
# Bumped from 2Gi 2026-08-18: OOMKilled repeatedly while recovering from
# a network disruption that forced several rapid restarts in a row -
# each restart's startup reconciliation is memory-hungry, and 2Gi wasn't
# enough headroom for that burst even though steady-state usage is
# normally fine. Nodes have ample free memory (26-36% used).
memory: 3Gi
# Allow Rancher to manage the local cluster
addLocal: "true"
# Audit logging
auditLog:
level: 1
maxAge: 7
maxBackup: 3
maxSize: 100
# Bootstrap password - will be rotated post-deploy via Authentik SSO
# This is used only for initial access before SSO is configured
bootstrapPassword: "ChangeMe-BootstrapOnly"
# Extra environment variables (not used for OIDC - configured via AuthConfig CRD post-deploy)
extraEnv: []
# Disable Rancher's bundled cert-manager (we already have it)
certmanager:
version: ""
# Feature flags
features: "multi-cluster-management=true"