メインコンテンツまでスキップ
バージョン: 3.6 (unsupported)

Components to Regularly Check When Running in a Kubernetes Environment

Most of the components deployed by manual deployment guides are self-healing with the help of the managed Kubernetes services and Kubernetes self-healing capability. There are also configured alerts that occur when some unexpected behavior happens. Thus, there shouldn't be so many things to do day by day for the deployment of Scalar products on the managed Kubernetes cluster. However, it is recommended to check the status of a system on a regular basis to see if everything is working fine. Here is the list of things you might want to do on a regular basis.

Kubernetes resources

Check if Pods are all healthy statues

Please check the Kubernetes namespaces:

  • default (or specified namespace when you deploy Scalar products) for the Scalar product deployment
  • monitoring for the Prometheus Operator and Loki

What to check:

  • STATUS is all Running
  • Pods are evenly distributed on the different nodes
$ kubectl get pod -o wide -n <namespace>
NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES
scalardb-7876f595bd-2jb28 1/1 Running 0 2m35s 10.244.2.6 k8s-worker2 <none> <none>
scalardb-7876f595bd-rfvk6 1/1 Running 0 2m35s 10.244.1.8 k8s-worker <none> <none>
scalardb-7876f595bd-xfkv4 1/1 Running 0 2m35s 10.244.3.8 k8s-worker3 <none> <none>
scalardb-envoy-84c475f77b-cflkn 1/1 Running 0 2m35s 10.244.1.7 k8s-worker <none> <none>
scalardb-envoy-84c475f77b-tzmc9 1/1 Running 0 2m35s 10.244.3.7 k8s-worker3 <none> <none>
scalardb-envoy-84c475f77b-vztqr 1/1 Running 0 2m35s 10.244.2.5 k8s-worker2 <none> <none>
$ kubectl get pod -n monitoring -o wide
NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES
alertmanager-scalar-monitoring-kube-pro-alertmanager-0 2/2 Running 1 (11m ago) 12m 10.244.2.4 k8s-worker2 <none> <none>
prometheus-scalar-monitoring-kube-pro-prometheus-0 2/2 Running 0 12m 10.244.1.5 k8s-worker <none> <none>
scalar-logging-loki-0 1/1 Running 0 13m 10.244.2.2 k8s-worker2 <none> <none>
scalar-logging-loki-promtail-2c4k9 0/1 Running 0 13m 10.244.0.5 k8s-control-plane <none> <none>
scalar-logging-loki-promtail-8r48b 1/1 Running 0 13m 10.244.3.2 k8s-worker3 <none> <none>
scalar-logging-loki-promtail-b26c6 1/1 Running 0 13m 10.244.2.3 k8s-worker2 <none> <none>
scalar-logging-loki-promtail-sks56 1/1 Running 0 13m 10.244.1.2 k8s-worker <none> <none>
scalar-monitoring-grafana-77c4dbdd85-4mrn7 3/3 Running 0 12m 10.244.3.4 k8s-worker3 <none> <none>
scalar-monitoring-kube-pro-operator-7575dd8bbd-bxhrc 1/1 Running 0 12m 10.244.1.3 k8s-worker <none> <none>

Check if Nodes are all healthy statuses

What to check:

  • STATUS is all Ready
$ kubectl get nodes
NAME STATUS ROLES AGE VERSION
k8s-control-plane Ready control-plane 16m v1.25.3
k8s-worker Ready <none> 15m v1.25.3
k8s-worker2 Ready <none> 15m v1.25.3
k8s-worker3 Ready <none> 15m v1.25.3

Prometheus dashboard (Alerts of Scalar products)

Access to the Prometheus dashboard according to the document Monitoring Scalar products on the Kubernetes cluster. In the Alerts tab, you can see the alert status.

What to check:

  • All alerts are green (Inactive)

If some issue is occurring, it shows you red (Firing) status.

Grafana dashboard (metrics of Scalar products)

Access to the Grafana dashboard according to the document Monitoring Scalar products on the Kubernetes cluster. In the Dashboards tab, you can see the dashboard of Scalar products. In these dashboards, you can see some metrics of Scalar products.

Those dashboards cannot address issues directly, but you can see changes from normal (e.g., increasing transaction errors) to get hints for investigating issues.