Prometheus: мониторинг AWS Load Balancer Controller Controller в AWS EKS
“Внезапно”, но AWS ALB Controller умеет в метрики для Prometheus. Используем стандартный URI /metrics. Для начала – проверим, что метрики в поде есть. Находим под: kk -n kube-system get pod | grep awsaws-load-balancer-controller-7bdcf75789-wk4bt 1/1 Running 0 41h Пробрасываем на него порт: kk port-forward -n kube-system pods/aws-load-balancer-controller-7bdcf75789-wk4bt 8080:8080Forwarding from 127.0.0.1:8080 -> 8080Forwarding from [::1]:8080 -> 8080 И пробуем с рабочей машины: curl -s localhost:8080/metrics | headHELP aws_api_call_duration_seconds Perceived latency from when your code makes an SDK call, includes retriesTYPE aws_api_call_duration_seconds histogramaws_api_call_duration_seconds_bucket{operation="AuthorizeSecurityGroupIngress",service="EC2",le="0.005"} 0aws_api_call_duration_seconds_bucket{operation="AuthorizeSecurityGroupIngress",service="EC2",le="0.01"} 0aws_api_call_duration_seconds_bucket{operation="AuthorizeSecurityGroupIngress",service="EC2",le="0.025"} 0aws_api_call_duration_seconds_bucket{operation="AuthorizeSecurityGroupIngress",service="EC2",le="0.05"} 0aws_api_call_duration_seconds_bucket{operation="AuthorizeSecurityGroupIngress",service="EC2",le="0.1"} 0aws_api_call_duration_seconds_bucket{operation="AuthorizeSecurityGroupIngress",service="EC2",le="0.25"} 29aws_api_call_duration_seconds_bucket{operation="AuthorizeSecurityGroupIngress",service="EC2",le="0.5"} 39aws_api_call_duration_seconds_bucket{operation="AuthorizeSecurityGroupIngress",service="EC2",le="1"} 39 Супер – метрики есть. Осталось их собрать. Contents Настройка PrometheusALB Controller ServiceALB Controller ServiceMonitor Настройка Prometheus У нас…
https://rtfm.co.ua/prometheus-monitoring-aws-load-balancer-controller-controller-v-aws-eks/
Post #278
562