Content
# Pingera MCP Server
A Model Context Protocol (MCP) server for the [Pingera monitoring service](https://pingera.ru?utm_source=mcp_readme), providing seamless integration between AI models and monitoring data.
## Features
- **Modular Architecture**: Separate Pingera API client library with clean abstractions
- **Flexible Operation Modes**: Run in read-only or read-write mode
- **MCP Tools**: Execute monitoring operations through tools (list_pages, get_page_details, test_connection)
- **Robust Error Handling**: Comprehensive error handling with custom exception hierarchy
- **Real-time Data**: Direct integration with Pingera API v1 for live monitoring data
- **Type Safety**: Pydantic models for data validation and serialization
- **Configurable**: Environment-based configuration management
## Quick Start
### Prerequisites
- Python 3.10+
- UV package manager
- Pingera API key - get one at [app.pingera.ru](https://app.pingera.ru)
### Installation and Setup
```bash
# Install dependencies
uv sync
# Set up your API key (required)
# Add PINGERA_API_KEY to your environment
# Run the server
python -m pingera_mcp
```
The server will start in read-only mode by default and connect to the Pingera API.
## Operation Modes
The Pingera MCP Server supports two transport modes:
### 1. Stdio Mode (Default)
Used for local integration with Claude Desktop and other stdio-based MCP clients. Authentication is configured via environment variables.
```bash
# Run in stdio mode (default)
python -m pingera_mcp
```
### 2. HTTP/SSE Mode
Enables web-based access through HTTP with Server-Sent Events. Perfect for remote clients, web applications, and programmatic access.
```bash
# Configure SSE mode
export TRANSPORT_MODE=sse
export HTTP_HOST=0.0.0.0
export HTTP_PORT=5000
# Run in SSE mode
python -m pingera_mcp
```
The server will be accessible at `http://0.0.0.0:5000/sse`
#### Authentication in SSE Mode
SSE mode supports two authentication methods via the `Authorization` header:
1. **API Key Authentication** (recommended for testing):
```
Authorization: YOUR_API_KEY
```
2. **JWT Bearer Token Authentication**:
```
Authorization: Bearer YOUR_JWT_TOKEN
```
#### Selective Tool Exposure
You can control which tools are available to clients using the `X-MCP-Tools` header in SSE mode. This is useful for limiting AI agents to specific operations or creating different access levels.
**Example: Limiting to read-only page tools**:
```python
client = MultiServerMCPClient({
"pingera": {
"url": "http://0.0.0.0:5000/sse",
"transport": "sse",
"headers": {
"Authorization": "YOUR_API_KEY",
"X-MCP-Tools": "list_pages,get_page_details,list_component_groups"
}
}
})
```
The `X-MCP-Tools` header accepts a comma-separated list of tool names. Only the specified tools will be exposed to the client. If the header is omitted, all tools compatible with the current server mode (read-only or read-write) will be available.
**Note**: Tool filtering respects the server's operation mode. Write-operation tools will not be exposed in read-only mode, even if requested via the header.
#### Example: Python Client for SSE Mode
```python
import asyncio
from langchain_mcp_adapters.client import MultiServerMCPClient
async def main():
client = MultiServerMCPClient({
"pingera": {
"url": "http://0.0.0.0:5000/sse",
"transport": "sse",
"headers": {
"Authorization": "YOUR_API_KEY", # or "Bearer YOUR_JWT_TOKEN"
}
}
})
# Get available tools
tools = await client.get_tools()
print(f"Available tools: {len(tools)}")
asyncio.run(main())
```
#### SSE Mode Configuration Options
- **`TRANSPORT_MODE`** - Set to `sse` to enable HTTP/SSE mode (default: `stdio`)
- **`HTTP_HOST`** - Host to bind to (default: `0.0.0.0`)
- **`HTTP_PORT`** - Port to listen on (default: `5000`)
- **`REQUIRE_AUTH_HEADER`** - Require Authorization header (default: `false`)
**Note**: In SSE mode, you can either:
- Provide credentials via environment variables (PINGERA_API_KEY or PINGERA_JWT_TOKEN) as fallback
- Send credentials in the Authorization header with each request (recommended)
When `REQUIRE_AUTH_HEADER=true`, the server will only accept requests with valid Authorization headers.
## Claude Desktop Integration
To use this MCP server with Claude Desktop, you need to configure it in your Claude Desktop settings.
### Installation
First, install the package globally using UV:
```bash
uv tool install pingera-mcp-server
```
### Configuration
Open the Claude Desktop configuration file:
- **macOS**: `~/Library/Application Support/Claude/claude_desktop_config.json`
- **Windows**: `%APPDATA%/Claude/claude_desktop_config.json`
Add the following configuration:
```json
{
"mcpServers": {
"pingera": {
"command": "uv",
"args": [
"run",
"--with",
"pingera-mcp-server",
"--python",
"3.10",
"python",
"-m",
"pingera_mcp"
],
"env": {
"PINGERA_API_KEY": "your_api_key_here",
"PINGERA_MODE": "read_only",
"PINGERA_BASE_URL": "https://api.pingera.ru/v1",
"PINGERA_TIMEOUT": "30",
"PINGERA_MAX_RETRIES": "3",
"PINGERA_DEBUG": "false",
"PINGERA_SERVER_NAME": "Pingera MCP Server"
}
}
}
}
```
### Required Environment Variables
- **`PINGERA_API_KEY`** - Your Pingera API key (required)
### Optional Environment Variables
- **`PINGERA_MODE`** - Operation mode: `read_only` (default) or `read_write`
- **`PINGERA_BASE_URL`** - API endpoint (default: `https://api.pingera.ru/v1`)
- **`PINGERA_TIMEOUT`** - Request timeout in seconds (default: `30`)
- **`PINGERA_MAX_RETRIES`** - Maximum retry attempts (default: `3`)
- **`PINGERA_DEBUG`** - Enable debug logging (default: `false`)
- **`PINGERA_SERVER_NAME`** - Server display name (default: `Pingera MCP Server`)
### Restart Claude Desktop
After updating the configuration file, restart Claude Desktop to load the new MCP server. You should now be able to access your Pingera monitoring data directly through Claude's interface.
### Verify Installation
Once configured, you can ask Claude to:
- "What are my status pages and their names?"
- "Show details of a page with id ..."
- "When was the last failed check for check ..."
- "Run a simple synthetic check for the website with URL https://"
## Configuration
Configure the server using environment variables:
```bash
# Required (for stdio mode or as fallback in SSE mode)
PINGERA_API_KEY=your_api_key_here
# OR
PINGERA_JWT_TOKEN=your_jwt_token_here
# Transport Mode
TRANSPORT_MODE=stdio # stdio (default) or sse
# SSE Mode Configuration (only used when TRANSPORT_MODE=sse)
HTTP_HOST=0.0.0.0 # Host to bind to
HTTP_PORT=5000 # Port to listen on
REQUIRE_AUTH_HEADER=false # Require Authorization header
# Optional
PINGERA_MODE=read_only # read_only or read_write
PINGERA_BASE_URL=https://api.pingera.ru/v1
PINGERA_TIMEOUT=30
PINGERA_MAX_RETRIES=3
PINGERA_DEBUG=false
PINGERA_SERVER_NAME=Pingera MCP Server
```
## MCP Tools
Available tools for AI agents:
### Pages Management
- **`list_pages`** - Get paginated list of monitored pages
- Parameters: `page`, `per_page`, `status`
- **`get_page_details`** - Get detailed information about a specific page
- Parameters: `page_id`
### Component Management
- **`list_component_groups`** - List all component groups for monitoring organization
- **`get_component_details`** - Get detailed information about a specific component
- Parameters: `component_id`
### Monitoring Checks
- **`list_checks`** - List all monitoring checks (HTTP, TCP, ping, etc.)
- Parameters: `page`, `page_size`, `status`, `check_type`
- **`get_check_details`** - Get detailed information about a specific check
- Parameters: `check_id`
### Alert Rules
- **`list_alert_rules`** - List all alert rules and their trigger conditions
### Heartbeat Monitoring
- **`list_heartbeats`** - List all heartbeat monitors for cron jobs and scheduled tasks
- Parameters: `page`, `page_size`, `status`
### Incident Management
- **`list_incidents`** - List all incidents and their current status
- Parameters: `page`, `page_size`, `status`
### Connection Testing
- **`test_pingera_connection`** - Test API connectivity
### Write Operations
Available only in read-write mode (`PINGERA_MODE=read_write`):
#### Pages Management
- **`create_page`** - Create a new status page
- Parameters: `name` (required), `subdomain`, `domain`, `url`, `language`
- **`update_page`** - Update existing status page configuration
- Parameters: `page_id` (required), `name`, `subdomain`, `domain`, `url`, `language`, additional kwargs
- **`patch_page`** - Partially update specific page fields
- Parameters: `page_id` (required), kwargs for specific fields
- **`delete_page`** - Permanently delete a status page
- Parameters: `page_id` (required)
#### Component Management
- **`create_component`** - Create new component or component group
- Parameters: `page_id` (required), `name` (required), `description`, `group`, `group_id`, `only_show_if_degraded`, `position`, `showcase`, `status`
- **`update_component`** - Update existing component configuration
- Parameters: `page_id` (required), `component_id` (required), `name`, `description`, `group`, `group_id`, `only_show_if_degraded`, `position`, `showcase`, `status`, additional kwargs
- **`patch_component`** - Partially update specific component fields
- Parameters: `page_id` (required), `component_id` (required), kwargs for specific fields
- **`delete_component`** - Delete a component permanently
- Parameters: `page_id` (required), `component_id` (required)
#### Monitoring Checks
- **`create_check`** - Create new monitoring check
- Parameters: `check_data` (dict with check configuration)
- **`update_check`** - Update existing monitoring check
- Parameters: `check_id` (required), `check_data` (dict with updated configuration)
- **`delete_check`** - Delete monitoring check permanently
- Parameters: `check_id` (required)
- **`pause_check`** - Temporarily pause monitoring check
- Parameters: `check_id` (required)
- **`resume_check`** - Resume paused monitoring check
- Parameters: `check_id` (required)
#### Alert Rules
- **`create_alert`** - Create new alert rule
- Parameters: `alert_data` (dict with alert configuration)
- **`update_alert`** - Update existing alert rule
- Parameters: `alert_id` (required), `alert_data` (dict with updated configuration)
- **`delete_alert`** - Delete alert rule permanently
- Parameters: `alert_id` (required)
#### Heartbeat Management
- **`create_heartbeat`** - Create new heartbeat monitor
- Parameters: `heartbeat_data` (dict with heartbeat configuration)
- **`update_heartbeat`** - Update existing heartbeat monitor
- Parameters: `heartbeat_id` (required), `heartbeat_data` (dict with updated configuration)
- **`delete_heartbeat`** - Delete heartbeat monitor permanently
- Parameters: `heartbeat_id` (required)
- **`send_heartbeat_ping`** - Manually send ping to heartbeat
- Parameters: `heartbeat_id` (required)
#### Incident Management
- **`create_incident`** - Create new incident on status page
- Parameters: `page_id` (required), `incident_data` (dict with incident details)
- **`update_incident`** - Update existing incident details
- Parameters: `page_id` (required), `incident_id` (required), `incident_data` (dict with updated details)
- **`delete_incident`** - Delete incident permanently
- Parameters: `page_id` (required), `incident_id` (required)
- **`add_incident_update`** - Add status update to incident
- Parameters: `page_id` (required), `incident_id` (required), `update_data` (dict with update details)
- **`update_incident_update`** - Edit existing incident update
- Parameters: `page_id` (required), `incident_id` (required), `update_id` (required), `update_data` (dict with updated content)
- **`delete_incident_update`** - Delete specific incident update
- Parameters: `page_id` (required), `incident_id` (required), `update_id` (required)
## Access Control Modes
The server supports two levels of access control:
### Read-Only Mode (Default)
- Access monitoring data
- View status pages and their configurations
- Test API connectivity
- No modification capabilities
### Read-Write Mode
- All read-only features
- Create, update and delete resources: status pages, checks, alerts, heartbeats
- Execute checks and get their results
- Manage incidents and notifications
Set `PINGERA_MODE=read_write` to enable write operations.
**Note**: Access control mode is independent of transport mode. You can run the server in read-write mode with either stdio or SSE transport.
## Architecture
### Pingera API Client Library
Located in `pingera/`, this modular library provides:
- **PingeraClient**: Main API client with authentication and error handling
- **Models**: Pydantic data models for type-safe API responses
- **Exceptions**: Custom exception hierarchy for error handling
### MCP Server Implementation
- **FastMCP Framework**: Modern MCP server implementation
- **Resource Management**: Structured access to monitoring data
- **Tool Registration**: Executable operations for AI agents
- **Configuration**: Environment-based settings management
## Testing
### Running the Test Suite
Run all tests:
```bash
uv run pytest
```
Run tests with verbose output:
```bash
uv run pytest -v
```
Run specific test files:
```bash
uv run pytest tests/test_models.py
uv run pytest tests/test_config.py
uv run pytest tests/test_mcp_server.py
```
Run tests with coverage:
```bash
uv run pytest --cov=pingera --cov=config --cov=mcp_server
```
### Test Structure
The test suite includes:
- **Unit Tests**: Testing individual components (models, config, client)
- **Integration Tests**: Testing MCP server functionality
- **Mock Tests**: Testing with simulated API responses
### Testing with MCP Inspector
The [MCP Inspector](https://github.com/modelcontextprotocol/inspector) is an official debugging tool that provides a web interface to test MCP servers interactively. It allows you to explore available resources, execute tools, and inspect the server's responses in real-time.
#### Setup Inspector
1. Create an `mcp.json` configuration file:
```json
{
"mcpServers": {
"pingera": {
"command": "uv",
"args": [
"run",
"--with",
"pingera-mcp-server",
"--python",
"3.10",
"python",
"-m",
"pingera_mcp"
],
"env": {
"PINGERA_API_KEY": "your_pingera_api_key",
"PINGERA_MODE": "read_only",
"PINGERA_BASE_URL": "https://api.pingera.ru/v1",
"PINGERA_TIMEOUT": "30",
"PINGERA_MAX_RETRIES": "3",
"PINGERA_DEBUG": "false"
}
}
}
}
```
2. Run the inspector:
```bash
npx @modelcontextprotocol/inspector --config mcp.json
```
3. Open your browser to the provided URL (typically `http://localhost:6274`)
#### Using Inspector
The inspector provides:
- **Resources Tab**: Browse available monitoring data resources
- **Tools Tab**: Execute MCP tools like `list_pages`, `get_page_details`, etc.
- **Logs Tab**: View detailed communication logs between the inspector and server
- **Interactive Testing**: Test tool parameters and see real-time responses
This is the recommended way to test your MCP server integration before deploying with Claude Desktop or other MCP clients.
### Manual Testing with `mcp_client.py`
`mcp_client.py` uses Gemini models and integrates with the MCP server to execute various tools. It is just an example and can be easly modified to use any other Large Language Model.
```bash
python mcp_client.py "Show me my status pages"
```
## Error Handling
The system includes comprehensive error handling:
- `PingeraError`: Base exception for all client errors
- `PingeraAPIError`: API response errors with status codes
- `PingeraAuthError`: Authentication failures
- `PingeraConnectionError`: Network connectivity issues
- `PingeraTimeoutError`: Request timeout handling
