Lab: Visualise Prometheus in Grafana (USE and RED methods)

What you will learn in this TP

• Provision a data source
• Apply USE and RED
• Make a dashboard dynamic

From query to dashboard #

What you will learn in this section

• Provision a data source
• Apply USE and RED
• Make a dashboard dynamic

Context. Metrics only help if the team sees them. You connect Grafana to Prometheus and build two panels following the reference methods: USE for resources, RED for services.

Prerequisites : Alertmanager lab  ·  Estimated time : 35 min

Data source, queries, variable.

1. 1. Provision the data source

   Why? Provisioning makes configuration reproducible: no clicks to redo after a reinstall.

   Add Grafana to the Compose file and declare Prometheus through provisioning.

2. 2. USE panel: CPU utilisation

   Why? USE (utilisation, saturation, errors) tells whether a resource is the bottleneck.

   Display the CPU percentage used per instance.

3. 3. RED panel: error rate

   Why? RED (rate, errors, duration) reflects what the user feels about a service.

   Compute the percentage of 5xx requests (use Prometheus's own metrics: prometheus_http_requests_total).

4. 4. Variable and import

   Why? Variables turn a dashboard into a tool reusable for every machine; importing avoids reinventing the wheel.

   Add an $instance variable then import the community dashboard "Node Exporter Full" (ID 1860).

Difficulty level : ●●○○○ (2/5)