Health status dashboards in Grafana let you monitor key infrastructure metrics across your Google Distributed Cloud (GDC) air-gapped appliance environment to identify potential issues. Viewing component health helps you diagnose unexpected system behavior, isolate root causes, and prioritize performance tuning or feature development.
GDC tracks and reports health status metrics for both hardware components, such as server nodes, and software resources, including Kubernetes clusters, VMs, and storage.
Before you begin
The root clusters contain a Grafana instance for infrastructure operators. These Grafana instances contain the health status dashboards.
You need role-based access controls to access data visualizations safely on dashboards of the Grafana instance. For this reason, follow the instructions of the Before you begin section to query and view metrics on dashboards to access the dashboards on the Grafana instances.
Grafana endpoint
Open one of the following URLs to access the Grafana endpoint of either the infra-obs project or the root admin cluster:
Grafana endpoint of the
infra-obsproject:https://GDC_URL/infra-obs/grafanaReplace GDC_URL with the URL of an organization in GDC.
Grafana endpoint of the root admin cluster:
https://ROOT_ADMIN_CLUSTER_URL/grafanaReplace ROOT_ADMIN_CLUSTER_URL with the URL of the root admin cluster in GDC.
Use case examples
This section includes examples on how you can use the dashboard for practical cases.
Performance tuning
If a component is performing poorly but not necessarily impacting SLO or firing alerts, you can communicate proactively to developers and prevent future issues from occurring.
Likewise, a component might want to know how its feature operates to make the right performance tradeoffs. Health status are one mechanism for collecting the information that components need.
Feature development
Suppose a customer requests changes or GDC plans to release a new feature. In that case, you can observe health status for the relevant components to determine the feasibility of supporting the new feature or change. Also, health status can be used to make product decisions when prioritizing the work.
For example, suppose a component has an average latency of 500 ms and is interested in reducing it to 250 ms. In that case, the team can calculate the relative cost of a 50% incremental reduction and compare it with designing an endpoint with 250 ms response time.