# Guance Cloud MCP

> Guance Cloud / 观测云 gives your AI agent full access to complex system monitoring data. Stop clicking through dashboards and log tabs. This MCP lets you ask natural language questions about infrastructure, retrieving monitor statuses, reviewing real-time events, and running deep data queries (DQL) without touching the Guance console.

## Overview
- **Category:** cloud-infrastructure
- **Price:** Free
- **Tags:** system-monitoring, incident-response, dashboarding, application-performance, infrastructure-metrics, real-time-alerts

## Description

Guance Cloud connects your AI agent directly to an entire observability stack. Forget logging into a dashboard just to check if a service is down or what caused the spike last night. This MCP lets you manage complex system monitoring, log analysis, and incident response simply by talking to your agent. You can ask it to list all monitors, pull detailed configuration reports on dashboards, or browse specific system events in real time. If you need deep metrics, you don't build a query; you just tell the agent what average CPU usage you want over the last hour. Your agent acts like an always-on site reliability assistant, keeping your infrastructure data accurate and giving you instant answers whether you’re troubleshooting or auditing resource usage. Connecting Guance Cloud to Vinkius means you get this power alongside thousands of other industry tools through one place.

## Tools

### get_billing
Retrieves your current cloud service usage and cost information.

### get_event
Fetches the detailed record for a specific observability event.

### get_monitor
Gets all configuration details for a single system monitor.

### get_workspace
Retrieves high-level metadata and status information about your monitoring workspace.

### list_access_keys
Shows a list of all API access keys configured in the workspace for auditing purposes.

### list_dashboards
Returns a comprehensive list of every dashboard configured in the system.

### list_events
Pulls a list of recent observability events, including alerts and errors.

### list_log_sources
Provides an inventory of all available log data sources for analysis.

### list_monitors
Lists every active and configured system monitor in the environment.

### query_data
Runs powerful Data Query Language (DQL) statements to fetch specific metrics or log data.

## Prompt Examples

**Prompt:** 
```
List all active monitors in Guance Cloud.
```

**Response:** 
```
I've retrieved your active monitors. You have 15 monitors configured, including 'High CPU Alert', 'API Latency', and 'Database Connection'. Would you like to check the status of the alerts?
```

**Prompt:** 
```
Show me recent events from the last hour.
```

**Response:** 
```
I've listed 5 events from the last hour. There was 1 Critical alert for 'Storage Full' and 4 Information events regarding system updates. Should I retrieve the details for the Critical alert?
```

**Prompt:** 
```
Query average CPU usage using DQL.
```

**Response:** 
```
I've executed the DQL query. The average CPU usage across your 'Production' cluster over the last 15 minutes is 42%. Would you like a breakdown by individual host?
```

## Capabilities

### Check system health status
The agent retrieves the current metadata and alert status for your entire monitoring workspace.

### List and check monitor details
You can ask the agent to list every configured system monitor and get its detailed setup information.

### Browse historical events
The agent pulls a stream of real-time observability records, including errors, alerts, and system changes.

### Run custom data queries
You tell the agent what metrics you need, and it executes powerful DQL statements to fetch specific logging or performance data.

### Audit infrastructure keys and billing
The agent provides visibility into your organizational API access keys and current usage billing records.

## Use Cases

### Troubleshooting an unexpected service dip
An SRE gets paged about high latency. Instead of jumping between the metrics dashboard and the event log, they ask their agent to list monitors and then execute a DQL query for average latency across the cluster last night. The agent answers with a clear number and identifies the time window where the issue started.

### Auditing resource access
An Infrastructure Lead needs to know which services are using which keys. They ask their agent to list access keys, check billing usage with `get_billing`, and confirm that all required dashboards are available via `list_dashboards`. It compiles a full compliance report instantly.

### Investigating a data loss incident
A technical analyst suspects log data is missing. They ask the agent to list log sources, then browse recent observability events using `list_events` to see if any deletion alerts fired. The agent pinpoints the exact time and source of the gap.

## Benefits

- Instead of manually running `list_monitors` to see what's broken, you simply ask your agent for the status. You get immediate visibility into every alert rule without leaving your chat window.
- Need to check past activity? Use the agent to browse real-time observability events via `list_events`. This means instantly knowing if a system change caused an error, which is way faster than digging through logs.
- The biggest time saver: running complex queries. Instead of building DQL in a separate tool and pasting it somewhere else, you just ask the agent to perform data querying using `query_data` directly.
- You don't have to guess what resources exist. Running `list_dashboards` gives you an immediate overview of your monitoring setup, letting you audit coverage instantly.
- It keeps everything connected: You can check the overall status with `get_workspace`, then drill down to specific alerts using `get_monitor`. It's a complete workflow built into one conversation.

## How It Works

The bottom line is you get instant answers about complex infrastructure metrics by talking to your agent instead of navigating web consoles.

1. Subscribe to this MCP and provide your Guance Cloud API Key (DF-API-KEY).
2. Connect the credentials to your preferred AI client, like Cursor or Claude.
3. Ask your agent a question—for example, 'What was the average latency last night?'—and it handles the data retrieval.

## Frequently Asked Questions

**How do I list monitors using Guance Cloud / 观测云 MCP?**
You ask your agent to execute `list_monitors`. The agent will return a full catalog of every monitor configured in the workspace, letting you know exactly what is being watched.

**Can I run custom metrics queries with Guance Cloud / 观测云 MCP?**
Yes. You use the `query_data` tool. Just tell your agent which metric and time window you want, and it runs the necessary DQL statement for you.

**What is the difference between list_events and get_event?**
`list_events` gives you a feed of recent occurrences (errors, alerts). `get_event` allows you to deep-dive into one specific event ID to see all its associated details.

**Do I need to manually find the API key for Guance Cloud / 观测云 MCP?**
The agent can help. You use `list_access_keys` to get a list of existing keys and confirm which credentials are available for your workspace.

**How do I check billing usage with Guance Cloud / 观测云 MCP?**
You ask the agent about billing, and it executes `get_billing`. You get an immediate breakdown of your current service consumption and costs.