Documentationbreadcrumb arrow Grafana Cloudbreadcrumb arrow Platformbreadcrumb arrow Assistantbreadcrumb arrow Get started
Grafana Cloud
Last reviewed: November 2, 2026

Get started with Grafana Assistant

Grafana Assistant is an AI-powered observability agent. Ask it questions in plain language, and it works across your metrics, logs, traces, profiles, and databases to give you answers without needing deep expertise in each tool.

Ask it things like:

  • “Show me CPU usage for the API service.”
  • “Create a dashboard for my database.”
  • “Why is my application responding slowly?”

Behind those questions, Assistant runs investigations, builds dashboards and queries, navigates your Grafana resources, and answers Grafana product questions. For the full picture, refer to What Assistant can do.

Get started in three steps

  1. Get a Grafana Cloud stack. Assistant works with your own data through a Grafana Cloud stack. If you don’t have one, create a free Grafana Cloud account.
  2. Set up Assistant. Choose Grafana Cloud or self-managed Grafana in Choose your setup.
  3. Start your first conversation. Open Assistant and ask a question in plain language.

Tip

Try Assistant in the public Grafana Play sandbox before you connect it to your own stack. Grafana Play includes sample data and lets you explore Assistant workflows without installing anything or signing in.

Choose your setup

Choose where Assistant runs:

  • Grafana Cloud: Choose this if you already use a Grafana Cloud stack or want the full feature set with no installation. Assistant runs directly in your stack.
  • Self-managed Grafana: Choose this if you run your own Grafana, such as OSS or Enterprise. You install the Assistant app and connect it to a Grafana Cloud stack, with a reduced feature set compared with Grafana Cloud.

Tip

Assistant includes a free tier. Refer to Grafana Assistant pricing for included seats, tiers, and limits.

What Assistant can do

After setup, Assistant acts as your observability expert across five core workflows:

  • Assistant Investigations: Ask questions about performance, launch investigations, and correlate metrics, logs, traces, profiles, and SQL data. For example: “Why did checkout error rates spike this afternoon?”
  • Manage dashboards: Create new monitoring dashboards or refine existing panels, layouts, and variables. For example: “Add a p95 latency panel to my API dashboard.”
  • Query data: Build and refine PromQL, LogQL, TraceQL, SQL, and k6 queries with validation and optimization tips. For example: “Write a LogQL query for 5xx errors in the last hour.”
  • Navigate Grafana resources: Find dashboards, data sources, and related tools without leaving the conversation. For example: “Find dashboards related to Kubernetes monitoring.”
  • Onboard and learn: Ask for Grafana guidance, observability concepts, monitoring strategies, and basic troubleshooting in context. For example: “Explain RED metrics and how to apply them to my services.”

Next steps

From here, you can: