-
Notifications
You must be signed in to change notification settings - Fork 2
McpServer
This page documents the Model Context Protocol (MCP) tools, prompts, and resources available in the CloudDebugger tool. These tools demonstrate the integration of MCP capabilities into this application.
Note: The MCP support in CloudDebugger is experimental. We welcome your feedback as we evaluate and improve this integration.
The MCP server endpoint is available at: /mcp/sse
--
The following MCP prompts have been added to CloudDebugger. These prompts enable analysis, health checks, and correlation of logs and traces using the Model Context Protocol.
-
Prompt:
analyze-trace <trace_id>
- Description: Analyze an OpenTelemetry trace for performance issues, bottlenecks, and failures.
-
Input:
-
trace_id
(string): The ID of the trace to analyze.
-
-
Output:
- Report covering:
- Trace overview and metrics
- Performance hotspots
- Error and failure analysis
- Bottleneck detection
- Optimization recommendations
- Report covering:
-
Prompt:
analyze-logs <minutes>
- Description: Analyze recent logs for errors, warnings, and performance patterns.
-
Input:
-
minutes
(int, default: 5): Number of minutes of logs to analyze.
-
-
Output:
- Log summary by level and category
- Error and warning pattern grouping
- Root cause analysis
- Remediation steps and prioritized action plan
-
Prompt:
correlate-trace-logs <trace_id>
- Description: Correlate a specific trace with related log entries for debugging.
-
Input:
-
trace_id
(string): The trace ID to correlate.
-
-
Output:
- Unified timeline of spans and log entries
- Error and performance correlation
- Debugging insights and recommendations
-
Prompt:
quick-health-check
- Description: Perform a quick health check of the CloudDebugger service based on recent telemetry.
-
Output:
- Current status (Healthy/Warning/Critical)
- Error rate (last 5 minutes)
- Average response time
- Top 3 active issues
- Immediate recommended actions
-
URI:
config://service/CloudDebugger
- Description: Returns the current configuration settings and parameters as seen by the application, excluding sensitive keys.
-
Output:
- All non-sensitive configuration key/values
- Provider information (source, type, values)
- Total settings and providers
-
URI:
env://service/CloudDebugger
- Description: Returns the current environment variables and system settings, filtering out sensitive data.
-
Output:
- All non-sensitive environment variables
- Total count
-
Latest Logs:
-
URI:
logs://service/CloudDebugger/app.log
Returns the latest 50 application log entries.
-
URI:
-
Logs by Trace ID:
-
URI:
logs://service/CloudDebugger/app.log/{traceId}
Returns all log entries associated with a specific trace ID.
-
URI:
-
Output:
- Timestamped log entries with level, category, message, exception info, trace/span IDs, and attributes
-
All Traces:
-
URI:
traces://service/CloudDebugger
Returns all available trace entries.
-
URI:
-
Trace by ID:
-
URI:
traces://service/CloudDebugger/{traceId}
Returns all trace entries for a specific trace ID.
-
URI:
-
Output:
- Trace metadata (IDs, operation, timing, status, tags)
The following tools allow secure file operations on the server:
Tool Name | Description | Parameters |
---|---|---|
list_files |
List all files and directories in a path |
path (string) |
read_file |
Read the contents of a text file |
path (string) |
write_file |
Create/overwrite a file with specified content |
path (string), content (string) |
delete_file |
Delete a file |
path (string) |
get_file_info |
Get info about a file or directory |
path (string) |
- All file tools return structured JSON with success/error, metadata, and details. Sensitive errors (e.g., access denied) are clearly indicated.
Tool Name | Description | Parameters |
---|---|---|
http_get |
Makes an HTTP GET request and returns the response |
url (string) |
-
Output:
- HTTP status, headers, content, response time, and error details if applicable.
- Security: Sensitive configuration and environment variables are automatically filtered.
- Error Handling: All tools return structured errors with type and message.
- Extensibility: Prompts and tools are designed for easy extension; see source code for annotation usage.
For further details, consult the source code or open an issue on the repository.