Skip to main content
So much telemetry goes unused without effective visualization. Grafana turns Prometheus metrics into dashboards that make platform health visible at a glance. In this lesson we’ll build a compact “Platform Dashboard” with four panel types—each chosen for a specific class of metric.
Before you begin, make sure a Prometheus data source is configured in Grafana. If you need setup guidance, see the Grafana data source docs.
Start by opening Grafana, then go to the left-hand menu and click Dashboards.
The image shows the settings for configuring a Prometheus data source in Grafana, including options for performance, query parameters, and HTTP methods.
Create a new dashboard and add a visualization.
The image shows the Grafana interface where a user can start building a new dashboard by adding visualizations, importing panels, or importing dashboards. The left sidebar contains navigation options such as Home, Bookmarks, and Dashboards.
Below are step-by-step instructions for adding four common panels: Time series, Stat, Gauge, and Pie chart. Follow the sequence to build an actionable dashboard.
  1. Time series — Request rate per pod
  • Use a Time series panel for trend analysis and per-entity lines.
  • Select the Time series visualization and choose your Prometheus data source.
  • Enter this PromQL to show request rate per pod over the last 5 minutes. sum by (pod) yields one line per pod:
  • Set the panel title to “Request Rate by Pod”.
  • Adjust the dashboard time range (e.g., last 30 minutes) so the chart reflects the desired window.
  • Save the panel and then save the dashboard as “Platform Dashboard”.
This helps you identify which pods receive the most traffic and whether load is well-distributed.
The image shows a Grafana interface with a dashboard editor. It includes sections for data queries, visualization options, and a panel displaying "No data."
  1. Stat — Errors in the last hour
  • Use a Stat panel when you need one prominent current value.
  • Query total application errors over the past hour with:
  • Title the panel “Errors — Last Hour”.
  • Add thresholds to color the stat (for example, green/yellow/red) and show a sparkline if you want a mini trend.
  • Save the panel to include it on the dashboard.
Use increase() when you want total counts over a window (e.g., errors during the last hour). Use rate() for per-second throughput or when plotting time-series rates.
  1. Gauge — Memory Used (%)
  • Gauges are ideal for capacity metrics and clear thresholding.
  • Compute used memory percentage with this PromQL:
  • Title the panel “Memory Used (%)” or a descriptive variant.
  • Configure thresholds, for example:
    • 0–70: green
    • 70–90: yellow
    • 90+: red
  • Save the panel. The gauge needle changes color as thresholds are crossed, giving an immediate view of capacity pressure.
The image shows a Grafana dashboard with metrics, including a gauge displaying the value 23.9, an error count of 955, and a graph of request rates by pod over time.
  1. Pie chart — Request distribution by HTTP status
  • Use a Pie chart for proportional breakdowns across categories.
  • Use increase() so each slice represents total counts during the chosen window (not instant rates):
  • Title the panel (for example, “Requests by HTTP Status — Last Hour”).
  • Enable legend and show values so slices are clearly labeled with counts and percentages.
  • Save the panel and refresh the dashboard.
The image shows a Grafana dashboard with a pie chart, a memory gauge, and an error count display.
Summary — What these panels cover
  • Trend analysis: Time series (request patterns over time).
  • Immediate health: Stat (current error count).
  • Thresholded resource usage: Gauge (memory utilization with thresholds).
  • Distribution: Pie chart (categorical splits like HTTP status codes).
Panel reference table Further reading and references Save your dashboard once more to commit all panels. You now have a concise Platform Dashboard that supports troubleshooting, capacity planning, and traffic analysis.

Watch Video

Practice Lab