# Honeybadger MCP

> Honeybadger MCP lets you monitor application health, track errors, and check uptime without leaving your development environment. Connect it to any AI client to list projects, analyze error groups, see backtraces from individual failures, or audit recent deployments using natural conversation.

## Overview
- **Category:** developer-tools
- **Price:** Free
- **Tags:** error-tracking, uptime-monitoring, exception-handling, application-health, cron-monitoring

## Description

Need to figure out why the staging site is glitching? Instead of jumping through five different dashboards—checking logs here, looking at deployment history there, and cross-referencing team ownership somewhere else—just talk to your agent. This MCP connects directly to Honeybadger, giving you full visibility into your app's health status using only conversation.

Your agent can list all monitored projects for you, then drill down to query fault groups to understand class names and environment distribution across your entire setup. If a specific error group looks bad, you can ask it to pull details from individual error notices, pulling backtraces and request data instantly. You can also check recent deployments or even audit team roles responsible for different projects. It’s powerful stuff, and since Vinkius hosts this MCP, connecting it is simple: just link your preferred AI client and start asking questions.

## Tools

### list_projects
Lists all projects in Honeybadger, returning key details like the project name, ID, language used, and current error counts.

### get_project
Retrieves complete information for a specific Honeybadger project when you provide its ID.

### list_faults
Lists error groups (faults) within a project, detailing class names, common messages, and how many times they occurred.

### get_fault
Gathers all details about one specific fault group, including its full history and scope.

### resolve_fault
Marks a specified Honeybadger fault as resolved, removing it from the list of active issues.

### list_notices
Lists individual error occurrences (notices) belonging to a specific fault group for review.

### get_notice
Gets the full, detailed record of one single error notice, including backtraces and session context.

### list_sites
Lists all external uptime monitoring sites associated with a project so you can check service availability.

### list_members
Retrieves the full list of team members registered on a project, including their assigned roles.

### list_deployments
Shows the most recent code deployments for a project, detailing who pushed the change and when it happened.

## Prompt Examples

**Prompt:** 
```
List all unresolved faults in my 'production-backend' project
```

**Response:** 
```
I've found 3 unresolved faults in 'production-backend': 'NoMethodError: undefined method [] for nil:NilClass', 'Timeout::Error', and 'ActiveRecord::RecordNotFound'. Which one would you like to inspect?
```

**Prompt:** 
```
Show me the details for fault ID 123456
```

**Response:** 
```
Fault ID 123456 ('NoMethodError'): First seen 2 days ago, occurred 45 times in 'production'. Last noticed today at 10:30 AM. It's currently unassigned. Would you like to see the last notice details or the backtrace?
```

**Prompt:** 
```
List recent deployments for project ID 9876
```

**Response:** 
```
I've listed the 3 most recent deployments for project 9876: Revision 'a1b2c3d' by Alex (production, 2 hours ago), Revision 'e5f6g7h' by Sarah (staging, yesterday), and Revision 'i9j0k1l' by Alex (production, 3 days ago).
```

## Capabilities

### Check Project Status
List all monitored applications, showing basic details like language used or the count of unresolved faults.

### Diagnose Error Groups
Query aggregate error groups (faults) to understand common patterns, class names, and where they are occurring across your infrastructure.

### Inspect Specific Failures
Retrieve deep details on single error occurrences (notices), including backtraces, request data, and the server environment that failed.

### Manage Faults and Uptime
Mark specific errors as resolved or ignore them to keep your dashboards clean. You can also check site availability and track recent code deployments.

### Audit Team Members
List the team members assigned to a project, helping you understand who owns which part of the bug-fixing workflow.

## Use Cases

### The production bug needs immediate root cause analysis
A developer sees a spike in errors. They ask their agent to list faults, identify the top error group, and then use get_notice on that fault to retrieve the full backtrace and request data. They find it's related to a specific API call missing required parameters.

### Investigating service degradation after an update
An SRE notices uptime dipping. They use list_deployments to see what code was pushed 3 hours ago, then check the fault data against that deployment to confirm the new revision caused the regression.

### Auditing project ownership and responsibilities
An Engineering Lead needs to know who should fix a critical bug. They ask the agent to list_members for the specific project, confirming which team member was assigned responsibility for that module.

### Clearing up stale error reports
A developer confirms a previously reported fault is fixed in the latest release. Instead of logging into the dashboard, they simply ask their agent to resolve_fault, keeping the error count accurate and clean.

## Benefits

- Pinpoint failures faster: Instead of manually searching through logs, you can ask your agent to list notices or get a notice for the exact backtrace needed. This drastically cuts down debugging time.
- Understand code changes instantly: By listing deployments and checking site availability, you immediately know if a recent code push is responsible for an increase in faults.
- Simplify triage workflows: Use the 'resolve_fault' tool or similar actions to mark issues as resolved right from your chat window. Your team stays focused only on what needs active work.
- Get project context quickly: Need to know who owns a bug? List members gives you that ownership data, while list_projects shows the overall status of all monitored services.
- Deep dive into every failure: Don't just see an error message. Use 'get_notice' to retrieve session context and request data for granular understanding of what triggered the fault.

## How It Works

The bottom line is you get a single source of truth for all your application health data without switching tabs or writing complex API calls.

1. Subscribe to this MCP and enter your Honeybadger Personal Auth Token.
2. Your AI client connects to the service, making all error monitoring data available through natural conversation.
3. You ask your agent specific questions—like 'What was wrong with the checkout page yesterday?'—and it returns detailed, actionable answers.

## Frequently Asked Questions

**How do I check if the error was caused by a recent deployment using Honeybadger MCP?**
You can use list_deployments first. This shows the revision history and who pushed what. Then, you cross-reference those dates with your fault data to see if an increase in errors aligns with a specific code change.

**Can I resolve faults using the Honeybadger MCP?**
Yes, you can use 'resolve_fault' to mark issues as fixed. This is useful when your team confirms that a critical bug has been addressed and should no longer be tracked.

**What information does get_notice provide for debugging?**
get_notice gives you the most granular data point: the full backtrace, request parameters, session context, and the exact server environment where that single error occurred. It's essential for root cause analysis.

**Does Honeybadger MCP help with uptime monitoring?**
Yes, it includes tools like list_sites to monitor site availability across different external endpoints. This lets you confirm if the failure is system-wide or limited to a specific service.

**Do I need developer credentials for Honeybadger MCP?**
You must provide your personal Auth Token when setting up the connection. The agent uses this token to authenticate and pull all the necessary project data on your behalf.