Any agent anywhere: How to use Grafana Cloud with Claude, Cursor, and any other AI agent

AI agents are becoming part of everyday engineering workflows, helping teams write code, troubleshoot issues, and automate tasks. But to be truly useful in production, these agents need access to more than code. They also need observability data and context to understand what's happening across your systems, investigate problems, and help you take action.

At Grafana Labs, we believe that context should be available wherever you work. That’s one of the core ideas behind Grafana Cloud as the open agentic observability platform: the flexibility to work with your observability data through the AI agents you prefer—whether it’s Grafana Assistant, a coding agent in your IDE, or a tool you’ve built yourself—so you can investigate and troubleshoot issues without switching tools.

In this post, we'll explore how to bring Grafana Cloud into your agentic workflows using tools like the Grafana Cloud MCP server and gcx, with practical examples along the way. You can also check out our ObservabilityCON 2026 talk in the video below to learn more.

Three ways to use Grafana Cloud with AI agents

There are multiple ways to bring Grafana Cloud observability data and context into your AI workflows, depending on where and how you prefer to work:

• Work directly in Grafana Cloud: Use Grafana Assistant, the AI-powered assistant built into Grafana Cloud, to investigate issues and interact with your observability data.
• Connect your preferred AI agent: Use the Grafana Cloud MCP server to bring Grafana Cloud data into compatible tools such as Claude, ChatGPT, and Codex.
• Work with coding agents and scripts: Use gcx, the Grafana Cloud command-line tool, to give coding agents access to observability data directly from their working environment.

Whether you're working directly in Grafana Cloud with Grafana Assistant or using your own AI agents, you can choose the approach that works best for you. Let's look at how the Grafana Cloud MCP server and gcx make that possible outside of Grafana Cloud.

Investigate an alert from your phone with Claude and Grafana Cloud MCP server

Imagine you're waiting in line at a coffee shop when a Slack message arrives from your director of engineering. Something is wrong with the frontend, and they need your help.

Your laptop is still in your backpack, but you don't necessarily need it to start investigating. With Claude connected to the Grafana Cloud MCP server, you can pull out your phone and ask a simple question: "Why is the frontend slow? Is there anything firing?"

Claude can use Grafana Cloud data to answer your question directly in the conversation. In this example, it confirms that an alert is firing, shows a chart of the relevant signal, and identifies which pages are affected.

You can then ask a follow-up question: how is this affecting the health of the frontend, and what's causing the problem? Again, Claude can return a visual Grafana Cloud view directly in the conversation, giving you more context about the health of your frontend application and helping you determine where to investigate next.

All of this happens without leaving the conversation or opening your laptop. By the time you get your coffee, you've already confirmed the issue and started exploring its impact.

How it works: The Grafana Cloud MCP server and MCP Apps

This experience is powered by two capabilities: the Grafana Cloud MCP server and MCP Apps.

The Model Context Protocol (MCP) is an open standard that enables AI agents to connect to external tools and data sources. The Grafana Cloud MCP server makes your Grafana Cloud observability data available to compatible agents, allowing them to query alerts, dashboards, and telemetry to help answer questions and investigate issues.

That means you can interact with Grafana Cloud data from the AI tools you already use, rather than rely exclusively on the assistant built into your observability platform.

MCP Apps take this a step further by bringing interactive Grafana Cloud views directly into your AI conversations. Instead of returning only a text description of an issue, an agent can display relevant visualizations, such as firing alerts and frontend health, alongside the conversation.

These views are delivered by the Grafana Cloud MCP server and can appear in MCP clients that support them.

To learn more, see the Grafana Cloud MCP server documentation. 

Continue your investigation in Cursor with Grafana Cloud context and gcx

Now imagine you've grabbed your coffee and opened your laptop. You want to take a closer look at the frontend performance issue and get some help from your coding agent.

You navigate to Grafana Cloud Frontend Observability, where a dashboard shows elevated time to first byte (TTFB), one of the symptoms of the slowdown.

You could manually copy the relevant queries, filters, and time range into your coding agent and explain what you're looking at. But Grafana Cloud provides a simpler way to hand off that context.

From a panel on the dashboard, you can click a button that automatically gathers the relevant information, including the query, filters, and time range. Then you select your agent of choice—Cursor, in this example—and send the context directly to it.

Now Cursor has the same starting context you were using in Grafana Cloud. Instead of explaining the issue from scratch, you can immediately ask it to compare the affected pods with healthy ones. Cursor uses gcx to query Grafana Cloud and begin investigating the differences.

This lets you continue investigating in the environment where you're already working, with the relevant observability context carried over from Grafana Cloud.

How it works: Give coding agents access to Grafana Cloud with gcx

gcx is the Grafana Cloud command-line tool. It provides a way for developers and coding agents to interact with Grafana Cloud directly from a terminal. Because coding agents can execute shell commands, they can also use gcx to access observability data as part of their investigations.

gcx includes curated agent skills that help coding agents understand how to use Grafana Cloud data for common tasks, without requiring step-by-step instructions. Agents can also discover available commands and use them to explore different parts of a system, including services, databases, and frontend sessions.

For example, gcx provides commands to:

• Explore application services: Inspect services and their operations to understand application behavior.
• Investigate databases: List and inspect database instances as part of a performance investigation.
• Examine frontend activity: Access frontend application information and sessions to understand the end-user experience.

Here's an example of available commands, with arguments omitted:

Copy

gcx
├─ appo11y
│  ├─ services get
│  └─ services list-operations
├─ dbo11y
│  ├─ instances list
│  └─ instances get
└─ frontend
   ├─ apps list
   └─ sessions get

In our example, gcx gives Cursor a way to look beyond the original alert and query additional observability data to guide its investigation. And because gcx works from the command line, this approach isn't limited to Cursor. Other coding agents that can execute shell commands can use it too, making it possible to incorporate Grafana Cloud data into existing development and automation workflows.

To learn more and get started, visit the gcx repository on GitHub.

Bring it all together: Investigate issues in Slack with an open source agent

What if you'd rather investigate an issue without leaving Slack? You can also connect an open source agent like Hermes to Grafana Cloud using both gcx and the Grafana Cloud MCP server.

For example, imagine your team reports a drop in completed bank transfers, even though traffic looks normal. You can ask Hermes directly in Slack to investigate the issue, providing the affected application, time window, and relevant context.

Hermes can then use gcx commands and MCP tools to investigate the underlying requests, examine user sessions, and compare the results with an earlier time window. From there, it can report its findings back in the Slack thread, with supporting Grafana Cloud links, potential causes, and recommended next steps.

Wrapping up

With Grafana Cloud, you choose the agent, the model, the provider, and where you work. Grafana Cloud then gives your agent access to the observability context and tools it needs to help investigate issues, identify root causes, and take action.

Ready to try it yourself? Here are three ways to get started:

1. Start with Grafana Assistant to investigate and troubleshoot directly in Grafana Cloud.
2. Connect your preferred agent through the Grafana Cloud MCP server to bring observability data into compatible AI tools.
3. Explore gcx to bring Grafana Cloud into your coding agents, scripts, and terminal-based workflows.

Miss out on ObservabilityCON 2026? Check out our blog post recapping all the big announcements!