Skip to main content
SmoothGlue Logo
← Back to Operations Overview

Get started with Grafana

TutorialPublicInterface: Grafana (via Ground Control)~10 minutes

What you'll build​

A Grafana dashboard with a live Prometheus visualization — from a blank dashboard to a working panel querying platform metrics. By the end you'll know how to navigate Grafana, use the Metrics Explorer, and run your first query.

Before you begin​

  • A SmoothGlue platform with Ground Control enabled
  • Console access with SRE or OrgAdmin role

Sign in to Grafana​

Navigate to https://grafana.<your-domain>/login.

On a SmoothGlue platform, Grafana uses Keycloak SSO. Click Log in and authenticate with your platform SSO credentials — same username and password you use for the Console and GitLab. You will not see a separate Grafana username/password prompt on production platforms.

Grafana login screen

In the left sidebar, click Dashboards.

Navigate to Dashboards

Create a new Dashboard​

Click + Create Dashboard.

Create a new Dashboard

Add a visualization​

On the new dashboard, click + Add visualization.

Add a visualization

Select a data source​

Select Prometheus from the data source list. Prometheus is pre-configured as the default data source — it's collecting metrics from all platform components already.

tip

You'll also see Loki in this list — that's the log data source. For metrics, use Prometheus.

Select Prometheus as the data source

Open the Metrics Explorer​

In the query editor, click into the Metric field. A dropdown appears. Click Metrics explorer → to open a searchable list of every metric Prometheus is collecting.

Open the Metrics Explorer

Find and select a metric​

Search for node_cpu to see CPU-related node exporter metrics. Select node_cpu_online — this shows the number of CPUs that are online and being scheduled across your platform nodes.

Search for node_cpu_online in the Metrics Explorer

Run the query​

Click Run queries to execute the query against Prometheus.

Click Run queries

Grafana renders a time-series panel showing one line per CPU core across all scraped nodes.

node_cpu_online visualization

What's next​

You've run your first Prometheus query in Grafana. From here you can:

  • Name and save the dashboard — click the save icon (top right) to persist it
  • Add more panels — click + Add in the dashboard toolbar to add another visualization
  • Explore your app's metrics — if your app exposes a /metrics endpoint with the correct Kubernetes annotations, its metrics are already in Prometheus. Search for your app's metric names in the Metrics Explorer

Learn more: