This section describes the built-in dashboards and charts provided by kube-opex-analytics.
For each namespace discovered in the target Kubernetes cluster, this dashboard section displays hourly usage trends for CPU and memory resources during the last week (7 days). The aim of these charts is to help understand trends of actual resource usage by each namespace, but also globally thanks to the stacked-area charts.
The date filter can be used to zoom out/in on a specific time range.
These charts are based on data consolidated hourly thanks to sample metrics collected every five minutes from Kubernetes.
For each namespace discovered in the target Kubernetes cluster, this dashboard section displays hourly usage/requests efficiency trends for CPU and memory resources during the last week (7 days). The aim of these charts is to help understand how efficient resource requests set on Kubernetes workloads are, compared against the actual resource usage over time.
The date filter can be used to zoom out/in on a specific time range.
These charts are based on data consolidated hourly thanks to sample metrics collected every five minutes from Kubernetes.
For the different namespaces discovered in the Kubernetes cluster, these charts show daily cumulative usage for CPU and memory resources during the last 2 weeks.
These charts are based on data consolidated hourly thanks to sample metrics collected every five minutes from Kubernetes.
Depending on the selected accounting model, the values on these charts can be actual costs (CHARGE_BACK model), cumulative usage (sum of hourly consolidated usage, CUMULATIVE_RATIO model), or a percentage of the global cluster usage (CHARGE_BACK model, 100% means the global cluster usage).
For the different namespaces discovered in the Kubernetes cluster, these charts show monthly cumulative usage for CPU and memory resources during the last 12 months.
The charts are based on data consolidated hourly thanks to sample metrics collected every five minutes from Kubernetes.
Depending on the selected accounting model, the values on these charts can be actual costs (CHARGE_BACK model), cumulative usage (sum of hourly consolidated usage, CUMULATIVE_RATIO model), or a percentage of the global cluster usage (CHARGE_BACK model, 100% means the total cluster capacity).
For each node discovered in the Kubernetes cluster, this dashboard section displays the CPU and the memory resources currently consumed by running pods. The data are refreshed every five minutes.
Any chart provided by kube-opex-analytics can be exported, either as PNG image, CSV or JSON data files.
- Go to the target chart section.
- Click on the link
export, then select the target exportion format. This action shall download the artifact instantly.
In addition or alternatively to the built-in dashboards, it's also possible to use Grafana for visualization thanks to the Prometheus exporter natively enabled by kube-opex-analytics.