Skip to content

An Agentic Model Context Protocol (MCP) server for real-time DevOps observability. Integrates as a GitHub App to analyze Actions, Deployments, and PR health, delivering visual "State of the Org" reports directly through AI-native interfaces.

License

Notifications You must be signed in to change notification settings

tsviz/github-mcp-server-devops

Repository files navigation

ActionsPulse πŸš€

Available on MCP Registry Docker Image Container Registry License: MIT

Real-time GitHub Actions observability through VS Code with GitHub Copilot

DORA Metrics β€’ Cost Analysis β€’ CI/CD Health β€’ Security Compliance

ActionsPulse Dashboard

☝️ This is an actual interactive HTML dashboard generated by ActionsPulse β€” see full tour below


✨ What is ActionsPulse?

ActionsPulse is an Agentic MCP Server that brings DevOps observability directly into your IDE. Ask GitHub Copilot questions like:

πŸ’¬ "Show me our DORA metrics for the last 30 days"
πŸ’¬ "Which workflows are costing us the most?"
πŸ’¬ "Generate a compliance report for SOC2"
πŸ’¬ "What's our deployment frequency this month?"

And get interactive visual dashboards with actionable insights.

🎯 Features

Feature Organization Enterprise
πŸ“Š DORA Metrics βœ… βœ… Enhanced
⚑ Performance Analytics βœ… βœ…
πŸ’° Cost Optimization βœ… βœ… Cross-org
πŸƒ Runner Utilization βœ… Self-hosted βœ… All runners
πŸ‘₯ Team Productivity βœ… βœ…
πŸ”’ Compliance Reports βœ… (GHAS) βœ…
πŸ’Ύ Cache Analytics βœ… βœ…
πŸŽ“ Maturity Assessment βœ… βœ…

πŸ–ΌοΈ Dashboard Tour

ActionsPulse generates interactive HTML dashboards that you can open in any browser. Simply ask Copilot:

πŸ’¬ "Generate DevOps reports for my organization"

Or use the generate_devops_reports tool directly. Here's a complete tour of what you get:

πŸ“Š Dashboard Overview β€” Executive summary with all key metrics at a glance Dashboard Overview

The main dashboard provides:

  • 🎯 DevOps Maturity Score with visual gauge
  • πŸ“ˆ DORA Metrics summary (Deployment Frequency, Lead Time, Change Failure Rate, MTTR)
  • ⚑ CI/CD Pipeline health overview
  • πŸ’° Cost analysis highlights
  • πŸ”’ Security & compliance status
  • Quick navigation to detailed reports
πŸ“ˆ DORA Metrics β€” Industry-standard DevOps performance indicators DORA Metrics Dashboard

Track the four key DORA metrics:

  • Deployment Frequency β€” How often you ship to production
  • Lead Time for Changes β€” Time from commit to production
  • Change Failure Rate β€” Percentage of deployments causing failures
  • Mean Time to Restore β€” How quickly you recover from incidents

Each metric includes trend analysis and benchmarks against industry standards (Elite, High, Medium, Low performers).

⚑ CI/CD Pipeline Health β€” Workflow performance and reliability CI/CD Health Dashboard

Deep dive into your CI/CD pipelines:

  • πŸ”„ Workflow success rates and trends
  • ⏱️ Average run times with P95/P99 latencies
  • 🚨 Failure analysis and common error patterns
  • πŸ“Š Per-repository and per-workflow breakdowns
  • πŸƒ Queue times and runner utilization
πŸ’° Cost Optimization β€” Runner costs and savings opportunities Cost Optimization Dashboard

Understand and optimize your GitHub Actions spending:

  • πŸ’΅ Total costs by runner type (GitHub-hosted vs self-hosted)
  • πŸ“Š Cost breakdown by repository, workflow, and OS
  • 🎯 Actionable recommendations for cost savings
  • πŸ“ˆ Spending trends and projections
  • ⚑ Efficiency metrics (cost per workflow, cost per minute)
πŸ”’ Security & Compliance β€” Audit-ready compliance reports Security Dashboard

Stay compliant and secure:

  • βœ… Compliance framework coverage (SOC2, ISO27001, HIPAA, PCI-DSS)
  • πŸ” Secret scanning status and alerts
  • πŸ›‘οΈ Code scanning findings
  • πŸ“‹ Branch protection rule compliance
  • πŸ” GHAS (GitHub Advanced Security) feature adoption
πŸŽ“ DevOps Maturity β€” Organizational capability assessment DevOps Maturity Dashboard

Assess your DevOps maturity level:

  • πŸ“Š Overall maturity score with visual gauge
  • 🎯 Category scores (CI/CD, Testing, Security, Monitoring, etc.)
  • πŸ“ˆ Improvement recommendations prioritized by impact
  • πŸ† Benchmark against industry standards
  • πŸ“‹ Actionable roadmap for advancement

πŸ’‘ How to generate these dashboards:

# Ask Copilot in natural language:
"Generate DevOps reports for the last 30 days"

# Or invoke the tool directly:
#generate_devops_reports --timeframe 30d

Reports are saved to timestamped folders as standalone HTML files β€” share them with your team or embed in internal wikis!


πŸš€ Quick Start

Install from MCP Registry

ActionsPulse is published to the official MCP Registry as io.github.tsviz/actions-pulse.

🌐 Web UI (Recommended) β€” Use the MCP Registry UI for 1-click installation into VS Code or Cursor:

  1. Go to vemonet.github.io/mcp-registry
  2. Search for "actions-pulse"
  3. Click Install β†’ Select your client (VS Code, Cursor)
  4. Configure environment variables when prompted

Manual Docker Setup β€” Or follow the steps below to configure manually.

Prerequisites

  • βœ… Docker installed
  • βœ… GitHub Personal Access Token (fine-grained recommended)
  • βœ… VS Code with GitHub Copilot

1. Create a Fine-Grained Personal Access Token

  1. Go to GitHub Settings β†’ Developer Settings β†’ Personal Access Tokens β†’ Fine-grained tokens

  2. Click Generate new token

  3. Configure basic settings:

    • Token name: actions-pulse-mcp
    • Expiration: 90 days (or per your security policy)
    • Resource owner: Select your organization
    • Repository access: All repositories
  4. Set Repository permissions:

    Permission Access Required Purpose
    Actions Read βœ… Yes Workflow runs, cache usage
    Administration Read βœ… Yes Billing data, repo settings
    Contents Read βœ… Yes Read config files from devops-config repo
    Custom properties Read βœ… Yes Read custom property values on repositories
    Deployments Read βœ… Yes Deployment frequency, environments (DORA)
    Discussions Read πŸ”Ά Optional Community engagement metrics
    Environments Read βœ… Yes Environment protection rules
    Issues Read βœ… Yes Issue metrics, resolution times (DORA)
    Merge queues Read πŸ”Ά Optional Merge queue adoption and wait times
    Metadata Read βœ… Yes Basic repo info (auto-granted)
    Pull requests Read βœ… Yes PR metrics, lead time, review times (DORA)
  5. Set Organization permissions:

    Permission Access Required Purpose
    Custom properties Read βœ… Yes Read property definitions/schemas at org level
    Custom properties for organizations Read βœ… Yes Read property values assigned to repositories
  6. Set Organization permissions (continued):

    Permission Access Required Purpose
    Members Read πŸ”Ά Optional Team membership for productivity metrics
    Self-hosted runners Read πŸ”Ά Optional Runner utilization metrics
    Administration Read βœ… Yes Org billing and settings
  7. Optional permissions (for compliance features, requires GitHub Advanced Security):

    Permission Access Required Purpose
    Secret scanning alerts Read ❌ Optional Compliance audit reports
    Code scanning alerts Read ❌ Optional Compliance audit reports
  8. Click Generate token and save it securely

2. Configure MCP Server

Option A: Using env-file (Recommended)

Add to your ~/.mcp.env:

GITHUB_TOKEN=ghp_your_fine_grained_token_here
πŸ“„ mcp.json with env-file

Add to VS Code's MCP settings (~/.vscode/mcp.json or workspace .vscode/mcp.json):

{
  "servers": {
    "actions-pulse": {
      "command": "docker",
      "args": [
        "run", "-i", "--rm",
        "--env-file", "/path/to/.mcp.env",
        "-e", "GITHUB_ORG=your-org-name",
        "ghcr.io/tsviz/actions-pulse:latest"
      ],
      "type": "stdio"
    }
  }
}

Option B: Direct Environment Variables

πŸ“„ mcp.json with inline env vars
{
  "servers": {
    "actions-pulse": {
      "command": "docker",
      "args": [
        "run", "-i", "--rm",
        "-e", "GITHUB_TOKEN=ghp_your_token",
        "-e", "GITHUB_ORG=your-org-name",
        "ghcr.io/tsviz/actions-pulse:latest"
      ],
      "type": "stdio"
    }
  }
}

3. Environment Variables Reference

Variable Required Description
GITHUB_TOKEN βœ… Yes Personal Access Token (fine-grained recommended)
GITHUB_ORG βœ… Yes Target GitHub organization to monitor (e.g., my-company). All API calls use this org.
DEFAULT_REPO_FILTER ❌ No Comma-separated list of repos to monitor (e.g., my-app,my-api). See precedence rules below.
GITHUB_API_URL ❌ No Custom API URL (default: https://api.github.com)
GITHUB_ENTERPRISE_SLUG ❌ No Enterprise slug for enhanced features
GITHUB_ENTERPRISE_URL ❌ No GitHub Enterprise Server API URL
DEVOPS_CONFIG_REPO ❌ No Config repo name (default: devops-config)
DEVOPS_CONFIG_PATH ❌ No Local path to config files (for mounted configs)

Repo Filter Precedence

When determining which repositories to query, ActionsPulse uses this precedence (highest to lowest):

Priority Source Applies To Example
1️⃣ repo_filter parameter in tool call Individual tools get_dora_metrics(repo_filter: "app1,app2")
2️⃣ inventory.yaml repositories generate_devops_reports Repos defined in config file
3️⃣ DEFAULT_REPO_FILTER env var All tools (fallback) DEFAULT_REPO_FILTER=my-app,my-api
4️⃣ All org repos via GitHub API All tools (default if nothing set)

Tip: For quick setup without a config repo, just set DEFAULT_REPO_FILTER in the MCP Registry installer. For richer metadata (team, tier, compliance tags), use inventory.yaml.

4. Configuration Files (Optional)

By default, ActionsPulse queries ALL repositories in your organization via the GitHub API β€” no configuration files are required. You can filter repos dynamically using tool parameters like repo_filter.

The optional configuration files let you define persistent metadata (teams, tiers, compliance tags) for filtering and reporting. There are two approaches:

Option A: Remote Config Repository (Recommended for Teams)

Create a devops-config repository in your organization with the following structure:

devops-config/
β”œβ”€β”€ devops-config.yaml          # Main configuration
β”œβ”€β”€ repositories/
β”‚   └── inventory.yaml          # List of repos to monitor
β”œβ”€β”€ policies/
β”‚   β”œβ”€β”€ workflow-policies.yaml  # CI/CD standards
β”‚   └── security-policies.yaml  # Security requirements
└── dashboards/                 # Dashboard configs

The MCP server will automatically discover and load from {org}/devops-config repo.

Option B: Local Config Files (For Development/Testing)

Mount a local config directory into the Docker container:

πŸ“„ mcp.json with config volume
{
  "servers": {
    "actions-pulse": {
      "command": "docker",
      "args": [
        "run", "-i", "--rm",
        "--env-file", "/path/to/.mcp.env",
        "-e", "GITHUB_ORG=your-org-name",
        "-e", "DEVOPS_CONFIG_PATH=/app/config",
        "-v", "/path/to/your/config:/app/config:ro",
        "ghcr.io/tsviz/actions-pulse:latest"
      ],
      "type": "stdio"
    }
  }
}

Repository Inventory Example

πŸ“„ inventory.yaml

Create repositories/inventory.yaml to define which repos to monitor:

apiVersion: actions-pulse/v1
kind: RepositoryInventory
metadata:
  name: my-inventory
  version: "1.0.0"
  description: "Repositories to monitor"

spec:
  discovery:
    enabled: false  # Only monitor explicit repos

  repositories:
    - name: my-app
      team: platform
      tier: tier-1
      compliance: [SOC2]
      tags: [java, production]

    - name: my-api
      team: backend
      tier: tier-2
      tags: [nodejs, staging]

Repository Tiers Quick Reference

Tier Priority Uptime Response Time Use Case
tier-1 πŸ”΄ Critical 99.9% < 15 min Production, customer-facing
tier-2 🟑 Standard 99% < 1 hour Internal tools, staging
tier-3 🟒 Low Best effort < 24 hours Demos, prototypes

See docs/ARCHITECTURE.md for complete tier definitions, compliance requirements, and alerting behavior.

5. Restart VS Code

After updating mcp.json, restart VS Code to pick up the new MCP server. You can verify the server is running by opening GitHub Copilot Chat and asking about your DevOps metrics.

πŸ› οΈ Available Tools

πŸ“Š Usage & Performance Metrics

get_actions_usage_metrics

Analyze GitHub Actions usage and billing data (basic).

Parameters:
- org_name: Organization name (optional if GITHUB_ORG is set)
- timeframe: '24h' | '7d' | '30d'
- breakdown: 'repository' | 'workflow' | 'runner_type'

get_detailed_usage_metrics ⭐

GitHub Insights-style detailed usage metrics with per-workflow, per-job, per-repo, per-OS, and per-runner breakdowns.

Parameters:
- org_name: Organization name (optional if GITHUB_ORG is set)
- timeframe: '7d' | '30d' | '90d'
- repo_filter: Comma-separated list of repositories (optional)

get_detailed_performance_metrics ⭐

GitHub Insights-style performance metrics with avg run time, queue time, and failure rates per workflow/job/repo/OS/runner.

Parameters:
- org_name: Organization name (optional if GITHUB_ORG is set)
- timeframe: '7d' | '30d' | '90d'
- repo_filter: Comma-separated list of repositories (optional)

get_actions_performance_metrics

Get workflow performance analytics with P95/P99 latencies (basic).

Parameters:
- org_name: Organization name (optional if GITHUB_ORG is set)
- repo_name: Specific repository (optional)
- workflow_id: Specific workflow (optional)
- timeframe: '1h' | '6h' | '24h' | '7d'
πŸƒ Runners & Cost Optimization

Enhanced Cost Detection: Reports now use a three-tier system for accurate runner cost calculation:

  • 🎯 API Detection - Uses hosted runners API for exact machine specs
  • 🏷️ Label Detection - Pattern matching against runner catalog
  • πŸ“Š Default Pricing - OS-based fallback

See Configuration Guide for details.

analyze_runner_utilization

Analyze runner utilization and efficiency.

Parameters:
- org_name: Organization name (optional if GITHUB_ORG is set)
- runner_type: 'self-hosted' | 'github-hosted' | 'all'
- include_costs: Include cost analysis (default: true)

get_actions_cache_analytics

Analyze Actions cache usage and efficiency.

Parameters:
- org_name: Organization name (optional if GITHUB_ORG is set)
- repo_name: Specific repository (optional)
- timeframe: '24h' | '7d' | '30d'

generate_cost_optimization_report

Generate actionable cost optimization recommendations.

Parameters:
- org_name: Organization name (optional if GITHUB_ORG is set)
- include_recommendations: Include actionable recommendations (default: true)
- target_savings_percentage: Target savings (5-50, default: 20)
πŸ” Workflow Insights & Team Productivity

get_workflow_insights

Get workflow insights with bottleneck detection.

Parameters:
- org_name: Organization name (optional if GITHUB_ORG is set)
- repo_name: Repository name (required)
- workflow_name: Workflow name or filename (required)
- analyze_dependencies: Analyze job dependencies (default: true)

get_team_productivity_metrics

Analyze team productivity based on Actions and commit data.

Parameters:
- org_name: Organization name (optional if GITHUB_ORG is set)
- team_slug: Team slug (optional)
- include_individuals: Include individual metrics (default: false)
- timeframe: '7d' | '30d' | '90d'

get_compliance_audit_report

Generate compliance and security audit report.

Parameters:
- org_name: Organization name (optional if GITHUB_ORG is set)
- compliance_framework: 'SOC2' | 'ISO27001' | 'HIPAA' | 'PCI-DSS' | 'CUSTOM'
- include_secrets_scan: Include secret scanning (default: true, requires GHAS)

πŸ“Š DORA Metrics & Developer Experience

πŸ“ˆ DORA Metrics

get_dora_metrics

Get DORA metrics (Deployment Frequency, Lead Time, Change Failure Rate, Time to Restore).

Parameters:
- org_name: Organization name (optional if GITHUB_ORG is set)
- timeframe: '7d' | '30d' | '90d'
- repo_filter: Comma-separated list of repositories (optional)

get_enhanced_dora_metrics

DORA metrics using actual GitHub Deployments API for maximum accuracy.

Parameters:
- org_name: Organization name (optional if GITHUB_ORG is set)
- timeframe: '7d' | '30d' | '90d'
- repo_filter: Comma-separated list of repositories (optional)

get_pull_request_metrics

Pull request metrics including lead time, merge rates, and size distribution.

Parameters:
- org_name: Organization name (optional if GITHUB_ORG is set)
- timeframe: '7d' | '30d' | '90d'
- repo_name: Specific repository (optional)
- include_stale: Include stale PR analysis (optional)

get_issue_metrics

Issue metrics including time to close, label distribution, and backlog health.

Parameters:
- org_name: Organization name (optional if GITHUB_ORG is set)
- timeframe: '7d' | '30d' | '90d'
- repo_name: Specific repository (optional)
- label_filter: Filter by label (optional)

get_deployment_metrics

Deployment metrics from GitHub Deployments API.

Parameters:
- org_name: Organization name (optional if GITHUB_ORG is set)
- timeframe: '7d' | '30d' | '90d'
- environment: Filter by environment (optional)
- repo_filter: Comma-separated list of repositories (optional)

get_environment_metrics

Analyze GitHub environment configurations including protection rules.

Parameters:
- org_name: Organization name (optional if GITHUB_ORG is set)
- repo_filter: Comma-separated list of repositories (optional)

get_discussion_metrics

GitHub Discussions metrics including answer rates and engagement.

Parameters:
- org_name: Organization name (optional if GITHUB_ORG is set)
- repo_name: Specific repository (optional)
- timeframe: '7d' | '30d' | '90d'

get_merge_queue_metrics

Merge queue usage and adoption across repositories.

Parameters:
- org_name: Organization name (optional if GITHUB_ORG is set)
- repo_name: Specific repository (optional)

🏷️ Custom Properties

πŸ“‹ Custom Properties Tools

get_org_custom_properties

List all custom property definitions for an organization.

Parameters:
- org_name: Organization name (optional if GITHUB_ORG is set)

get_custom_properties_analytics

Analyze custom property usage and coverage across repositories.

Parameters:
- org_name: Organization name (optional if GITHUB_ORG is set)

get_repos_by_property

Find repositories by custom property value.

Parameters:
- org_name: Organization name (optional if GITHUB_ORG is set)
- property_name: Custom property name (e.g., team, tier, compliance)
- property_value: Property value to filter by (optional)

🏒 Enterprise Features (Optional)

βš™οΈ Enterprise configuration

If you have GitHub Enterprise, you can enable enhanced features by adding:

GITHUB_ENTERPRISE_SLUG=your-enterprise-slug

This enables:

  • Cross-organization billing aggregation
  • Enterprise-wide runner pools
  • Consolidated audit logs

πŸ”§ Development

πŸ› οΈ Build and run commands

Build locally

npm install
npm run build
docker build -t actions-pulse:local .

Run locally (without Docker)

export GITHUB_TOKEN=ghp_your_token
export GITHUB_ORG=your-org
npm start

πŸ€– Automating DevOps Reports with GitHub Actions

You can automate DevOps report generation using GitHub Copilot CLI with ActionsPulse MCP server in a GitHub Actions workflow. This enables scheduled weekly reports, on-demand analysis, and automatic issue creation with insights.

How It Works

  1. Install Copilot CLI in the workflow runner
  2. Configure ActionsPulse MCP with your GitHub token
  3. Run Copilot with a prompt to generate reports using MCP tools
  4. Create issues with the generated report

Example Workflow

See .github/workflows/weekly-devops-report.yml for a complete working example.

πŸ“„ Key workflow steps
- name: Setup MCP config
  env:
    GITHUB_TOKEN: ${{ secrets.GH_PAT_DEVOPS }}
  run: |
    mkdir -p ~/.copilot
    printf '%s\n' '{
      "mcpServers": {
        "actions-pulse": {
          "command": "docker",
          "args": ["run", "-i", "--rm", "-e", "GITHUB_TOKEN='"$GITHUB_TOKEN"'", "-e", "GITHUB_ORG=your-org", "ghcr.io/tsviz/actions-pulse:latest"],
          "tools": ["*"]
        }
      }
    }' > ~/.copilot/mcp-config.json

- name: Generate DevOps Report
  env:
    GITHUB_TOKEN: ${{ secrets.GH_PAT_DEVOPS }}
    GH_TOKEN: ${{ secrets.GH_PAT_DEVOPS }}
  run: |
    copilot --yolo \
      --disable-builtin-mcps \
      --additional-mcp-config @$HOME/.copilot/mcp-config.json \
      --prompt "Use the actions-pulse MCP tools to generate a DevOps report..."

Key Copilot CLI Flags

Flag Description
--yolo Auto-approve all tool calls (no confirmation prompts)
--disable-builtin-mcps Disable built-in MCP servers (use only custom ones)
--additional-mcp-config @<file> Load MCP server config from file (use $HOME not ~)
--prompt "<text>" The prompt for Copilot to execute

Tips

  • Use $HOME instead of ~ in the config path β€” tilde is not expanded in @file arguments
  • Embed the token in args β€” the env block in MCP config doesn't pass variables to Docker
  • Include "tools": ["*"] in your MCP config β€” this field is required

πŸ“š Documentation

Document Description
Quick Start Get up and running in 5 minutes
Configuration Guide Complete configuration reference
Architecture System design and tier definitions

Example Configurations

Ready-to-use configuration examples are available in the examples/ directory:

File Description
mcp-docker.json VS Code MCP config using Docker
mcp-local.json VS Code MCP config for local development
mcp-envfile.json VS Code MCP config using environment file
.env.example Environment variables template
inventory.yaml Repository inventory example
devops-config.yaml DevOps observer configuration
docker-compose.yml Docker Compose deployment

οΏ½πŸ“„ License

MIT

About

An Agentic Model Context Protocol (MCP) server for real-time DevOps observability. Integrates as a GitHub App to analyze Actions, Deployments, and PR health, delivering visual "State of the Org" reports directly through AI-native interfaces.

Resources

License

Stars

Watchers

Forks

Packages

 
 
 

Contributors 2

  •  
  •  

Languages