Monitoramento e Observabilidade

Stack de Observabilidade (kube-prometheus-stack)

# Instalar Prometheus + Grafana via Helm
helm repo add prometheus-community https://prometheus-community.github.io/helm-charts
helm upgrade --install kube-prometheus prometheus-community/kube-prometheus-stack \
  --namespace monitoring --create-namespace \
  --set grafana.adminPassword=admin

# Acessar
kubectl port-forward -n monitoring svc/kube-prometheus-grafana 3000:80
kubectl port-forward -n monitoring svc/kube-prometheus-prometheus 9090:9090

Métricas do Cluster

# Recursos padrão do kube-state-metrics:
# - kube_deployment_status_replicas
# - kube_pod_status_ready
# - kube_node_status_condition
# - container_cpu_usage_seconds_total
# - container_memory_working_set_bytes

# Query no PromQL
kubectl port-forward -n monitoring svc/prometheus-operated 9090
# PromQL: rate(container_cpu_usage_seconds_total[5m])
# PromQL: sum(container_memory_working_set_bytes) by (namespace)

Custom Metrics e HPA

apiVersion: autoscaling/v2
kind: HorizontalPodAutoscaler
metadata:
  name: api-hpa
spec:
  scaleTargetRef:
    apiVersion: apps/v1
    kind: Deployment
    name: api
  minReplicas: 2
  maxReplicas: 20
  metrics:
    - type: Resource
      resource:
        name: cpu
        target:
          type: Utilization
          averageUtilization: 70
    - type: Resource
      resource:
        name: memory
        target:
          type: Utilization
          averageUtilization: 80
    - type: Pods
      pods:
        metric:
          name: http_requests_per_second
        target:
          type: AverageValue
          averageValue: 1000

kubectl get hpa -w
kubectl describe hpa api-hpa

Loki — Logs Centralizados

# Helm install Loki (Simple Scalable)
helm upgrade --install loki grafana/loki \
  --namespace monitoring \
  --set deploymentMode=SingleBinary \
  --set loki.commonConfig.replication_factor=1

# Promtail — coleta logs dos nós
helm upgrade --install promtail grafana/promtail \
  --namespace monitoring \
  --set config.lokiAddress=http://loki:3100/loki/api/v1/push

# LogQL (Loki query language)
{namespace="prod", app="api"} |= "ERROR"
{namespace="prod"} |= "timeout" | json | duration > 5s
rate({app="nginx"} |= "404" [5m])

Alertmanager

# values-alertmanager.yaml
alertmanager:
  config:
    global:
      slack_api_url: 'https://hooks.slack.com/services/xxx'
    route:
      group_by: ['namespace', 'alertname']
      receiver: 'slack-prod'
      routes:
        - match:
            severity: critical
          receiver: 'pagerduty'
    receivers:
      - name: 'slack-prod'
        slack_configs:
          - channel: '#alerts'
            title: '{{ .GroupLabels.alertname }}'
            text: '{{ .CommonAnnotations.description }}'

helm upgrade kube-prometheus prometheus-community/kube-prometheus-stack \
  -f values-alertmanager.yaml

Custom Resource Definitions (CRDs)

# PodMonitor — coleta métricas de pods específicos
apiVersion: monitoring.coreos.com/v1
kind: PodMonitor
metadata:
  name: api-monitor
spec:
  selector:
    matchLabels:
      app: api
  podMetricsEndpoints:
    - port: metrics
      interval: 15s

kubectl get prometheus
kubectl get podmonitors
kubectl get servicemonitors

Observabilidade no K8s = Prometheus (métricas) + Loki (logs) + Grafana (dashboards) + Alertmanager (alertas). kube-prometheus-stack instala tudo de uma vez. PodMonitors/ServiceMonitors configuram coleta automática.