Labsco
ionhour logo

Ionhour

from ionhour

Let AI agents monitor and manage your infrastructure through the Model Context Protocol. Query, create, and resolve β€” all in natural language.

πŸ”₯πŸ”₯πŸ”₯βœ“ VerifiedAccount requiredNeeds API keys

@ionhour/mcp-server

Connect AI assistants to your IonHour uptime monitoring workspace using the Model Context Protocol.

Getting an API Key

  1. Go to your IonHour workspace Settings > API Keys
  2. Create a new key with Read & Write or Read Only permission
  3. Copy the key (starts with ionh_)

Available Tools (49)

Workspace

  • get_workspace - Get workspace details
  • whoami - Get API key info (workspace, user, permission level)
  • get_workspace_summary - Overview of projects, checks by status, and active incidents
  • get_workspace_reliability - Uptime, incident count, and MTTR over N days
  • list_team_members - List workspace members and roles
  • send_invitation - Invite users to the workspace

Projects

  • list_projects / create_project / update_project

Checks (Monitors)

  • register_check - Create a monitoring check (accepts human-readable intervals like "every 5 minutes" or "hourly")
  • list_checks / list_checks_by_status / find_check_by_name
  • get_check_status - Detailed status with recent signals
  • get_check_uptime - Uptime percentage with daily buckets
  • pause_check / resume_check
  • delete_check - Permanently delete a check and its data

Signals (Heartbeats)

  • send_heartbeat - Send a success signal
  • send_failure_signal - Report a failure
  • list_signals - View signal history

Incidents

  • list_incidents / search_incidents / get_incident
  • get_incident_timeline - Incident history for a check
  • create_incident / acknowledge_incident / resolve_incident
  • add_incident_note

Deployments

  • create_deployment - Start a deployment window (auto-pauses checks)
  • end_deployment - End deployment and resume checks
  • list_deployments

Dependencies

  • list_dependencies / get_dependency / create_dependency
  • update_dependency_status
  • delete_dependency - Delete a dependency

Status Pages

  • list_status_pages / create_status_page / update_status_page
  • create_announcement - Post a status update

Alert Channels

  • list_alert_channels / create_alert_channel / update_alert_channel
  • delete_alert_channel - Delete a notification channel

Escalation Rules

  • list_escalation_rules / create_escalation_rule / update_escalation_rule
  • delete_escalation_rule - Delete an escalation rule

Resources

The server exposes reference data that AI assistants can browse without calling tools:

URIDescription
ionhour://enumsAll IonHour enums (check statuses, incident states, severities, etc.)
ionhour://checks/schemaCheck creation schema with field constraints and status lifecycle
ionhour://help/ping-formatsHeartbeat integration examples (curl, Node, Python, cron, CI/CD)
ionhour://tools/catalogCatalog of all tools organized by domain with descriptions
ionhour://guides/workflowsCommon workflow patterns and best practices

Prompts

Pre-built workflow templates that guide AI assistants through multi-step operations:

PromptArgsDescription
diagnose_incidentincidentIdStep-by-step incident investigation
setup_monitoringserviceNameGuided new-service monitoring setup
deployment_checklistprojectIdPre/during/post-deploy workflow
weekly_reliability_reportdaysBack?Generate a reliability summary
triage_all_incidentsβ€”List and triage all active incidents
status_page_incidentincidentIdCommunicate an incident through status pages
dependency_health_auditβ€”Audit all dependencies and assess health impact

CLI Commands

npx @ionhour/mcp-server [command] [options]

Commands:
  (default)     Start the MCP server (stdio proxy mode)
  setup         Interactive wizard β€” install MCP server into your editors
  login         Authenticate via browser and store API key
  logout        Remove stored credentials
  whoami        Show current authentication status

Options:
  --api-key KEY     IonHour API key (or set IONHOUR_API_KEY env var)
  --base-url URL    MCP base URL (default: https://mcp.ionhour.com)
  --version, -v     Show version
  --help, -h        Show help

Environment Variables:
  IONHOUR_API_KEY   API key for authentication
  IONHOUR_BASE_URL  Base URL override
  IONHOUR_AUTH_URL  Keycloak auth server URL (for login)
  IONHOUR_REALM     Keycloak realm name (for login)

setup

Interactive wizard that detects installed editors (Cursor, Claude Desktop, VS Code, Claude Code, Windsurf) and writes the IonHour MCP server configuration into each one. Prompts for login if no credentials are stored:

npx @ionhour/mcp-server setup

login

Opens your browser to authenticate with IonHour, then automatically creates and stores an API key:

npx @ionhour/mcp-server login

Credentials are stored at ~/.config/ionhour/credentials.json (file permissions: 600). After login, you can run the MCP server without setting IONHOUR_API_KEY.

logout

Removes stored credentials from ~/.config/ionhour/credentials.json:

npx @ionhour/mcp-server logout

whoami

Shows the current authentication status β€” the workspace, user, and permission level associated with the stored API key:

npx @ionhour/mcp-server whoami

How It Works

This package runs a local MCP server over stdio that proxies requests to the IonHour API. Your AI assistant communicates with this local server, which forwards tool calls to your IonHour workspace.

AI Assistant <--stdio--> @ionhour/mcp-server <--HTTPS--> IonHour API

Documentation