Story Crater Bot
0af06b1239
- Loki log aggregation (MinIO backed, 10-day retention) - Promtail daemonset (pod + talos journal logs) - Prometheus + kube-state-metrics - Grafana dashboards (6-row template per service)
34 lines
975 B
YAML
34 lines
975 B
YAML
apiVersion: monitoring.coreos.com/v1
|
|
kind: PrometheusRule
|
|
metadata:
|
|
name: vault-rules
|
|
namespace: storage
|
|
spec:
|
|
groups:
|
|
- name: vault.rules
|
|
interval: 15s
|
|
rules:
|
|
- alert: VaultHighErrorRate
|
|
expr: |
|
|
(
|
|
sum(rate(vault_core_handle_request_total{error="true"}[5m]))
|
|
/
|
|
sum(rate(vault_core_handle_request_total[5m]))
|
|
) > 0.05
|
|
for: 10m
|
|
labels:
|
|
severity: warning
|
|
annotations:
|
|
summary: "High error rate on Vault"
|
|
description: "Error rate exceeded 5% for 10 minutes. Value: {{ $value | humanizePercentage }}"
|
|
|
|
- alert: VaultSealed
|
|
expr: |
|
|
vault_core_unsealed == 0
|
|
for: 1m
|
|
labels:
|
|
severity: critical
|
|
annotations:
|
|
summary: "Vault is sealed"
|
|
description: "Vault has been sealed for 1 minute. Immediate attention required."
|