M
MCP Directory CURATED DIRECTORY
911 MCP servers indexed
中文
← Back to directory
R

Rootly

Community
Rootly MCP Server: Incident management, on-call scheduling, and intelligent analysis
How to install ↓ Compare GitHub source repository ↗
Category
Dev Tools #60 of 438
Stars
★ 45 Popular
Transport
stdio (local process) · SSE · Streamable HTTP
Runtime
Python 3.12+ · Docker · Hosted (no local runtime)
Credentials
API key / credential required
License
Apache-2.0
Last commit
Tools
100
64FMRS · C

Rootly MCP Server is a powerful tool that provides comprehensive incident management, on-call scheduling, and intelligent analysis capabilities. It supports multiple MCP clients and offers a rich set of tools for managing various DevOps operations. With dynamic tool generation and intelligent analysis features, it helps teams respond to and manage incidents more effectively.

Strongest · Documentation 16/20 Weakest · Reliability 9/20

Reliability
9/20
Security and permissions
13/20
Maintenance
12/20
Documentation
16/20
Setup experience
14/20
Why each score
Reliability 9/20
The README provides a reasonably complete manifest structure (remotes with streamable-http/sse, a pypi stdio package) and documents `--list-tools` plus allowlist smoke-test commands, showing some attention to verifiability. However, no CI workflow or committed test suite is present in the reviewed material, so the claimed ~218/70 tool surfaces cannot be checked against real behavior from static evidence alone. More importantly, the docs contradict themselves: one example requires explicitly setting `ROOTLY_MCP_ENABLE_WRITE_TOOLS=true` to get write tools, while another section states hosted and self-hosted deployments now both expose the full tool surface by default — these two claims about default behavior conflict, undermining confidence that declared behavior matches actual behavior. Per the static calibration rule, reliability is capped at 12 absent executable verification evidence, so a 9 is warranted.
Security and permissions 13/20
Credential handling looks reasonable: remote auth uses a Bearer token flagged isSecret/isRequired in the manifest, the local package injects `ROOTLY_API_TOKEN` via environment variable, and OAuth2 auto-login is offered to avoid manual token handling. Scoping is partly good: API key management, user creation/deletion, role management, and webhook configuration are explicitly excluded, and delete operations are disabled by default — a sensible least-privilege boundary. Telemetry disclosure is fairly transparent, noting redaction of free-form event text and omission of actor names while retaining only stable internal IDs. However, the documentation is internally inconsistent about whether write tools (create_incident/update_* etc.) are enabled by default, which directly affects whether an agent could modify production incidents, schedules, or escalation policies without explicit opt-in — an incomplete scoping disclosure. This keeps the score out of the top band even though no red-line issue (no default irreversible delete/pay/remote-exec, no credential exfiltration, no fabricated identity or test results) was found.
Maintenance 12/20
The project uses the Apache-2.0 license and a PyPI version of 2.2.9, indicating multiple iterative releases; the repo is not archived and has only 4 open issues, suggesting modest but active usage. A CONTRIBUTING.md is referenced, implying some contribution process exists. That said, the README/manifest text alone cannot confirm actual commit cadence, release frequency, or issue-response times, and no explicit security-response channel or vulnerability-disclosure policy is documented, leaving a governance transparency gap.
Documentation 16/20
Documentation is extensive, covering hosted and local deployment, configuration examples for most mainstream clients (Claude Code, Cursor, Windsurf, Codex, Claude Desktop, Gemini CLI), tool profile switching (full/slim), workflow-focused tool subsets, Docker run examples, telemetry configuration with redaction details, and self-check commands (`--list-tools`, smoke tests). The main deduction is the internal contradiction about the default state of write tools (see reliability/security), which could mislead users about the default privilege scope, plus the absence of explicit rate-limit, cost, or SLA disclosure for the hosted API, and troubleshooting content limited mostly to tool-listing commands.
Setup experience 14/20
The hosted path has very low friction — just a URL plus OAuth2 or a bearer token — and the README supplies ready-to-copy configuration snippets for nearly every mainstream client. The local path via `uv`/`uvx` with the `ROOTLY_API_TOKEN` environment variable is clearly documented, and Docker examples cover all three transport modes. Per the static calibration rule, setup cannot exceed 15 without executable verification evidence (CI plus committed tests), so a 14 is given — reflecting clear steps that nonetheless depend entirely on the external hosted service (mcp.rootly.com) and documentation accuracy that this static review cannot independently confirm.

Static review · not runListed 2026-09-07

Read the FMRS scoring method →

Fit and risk

What it can accessUses the networkChanges third-party account data

Best for

  • DevOps teams
  • SRE engineers
  • Incident response teams
  • On-call managers
  • Platform engineers
  • Development teams needing incident management

Not for

  • Organizations requiring fully local deployment
  • Teams not using Rootly for incident management
  • Simple applications without need for real-time alerting and incident management

Required permissions

  • Requires Rootly API access
  • May need OAuth2 authentication or API token
  • May need write permissions for creating and updating incidents based on configuration

Risks and side effects

  • API token leakage risk
  • Dependency on Rootly service
  • Network connectivity issues may affect functionality
  • Large data requests may cause performance issues

Setup

Before you start

Runtime：Python 3.12+ · Docker · Hosted (no local runtime)

ROOTLY_API_TOKEN requiredsecret Required Rootly API token; generate it per the Rootly docs (docs.rootly.com). A Global API Key is recommended for organization-wide use.
ONCALLHEALTH_API_KEY optionalsecret On-Call Health (oncallhealth.ai) API key for check_oncall_health_risk to detect responder workload risk.
SENTRY_DSN optionalsecret Optional; exports AgentCat telemetry events to Sentry.
ROOTLY_OPENAI_APPS_CHALLENGE_TOKEN optionalsecret Optional; OpenAI Apps domain-verification token from the OpenAI submission portal.
Other optional settings (4)
ROOTLY_MCP_ENABLE_WRITE_TOOLS optional Set to true to enable write tools; the default surface is read-only.
ROOTLY_MCP_ENABLED_TOOLS optional Comma-separated allowlist of exact tool names to expose; fully replaces the default selection.
ROOTLY_MCP_HOSTED_TOOL_PROFILE optional Server-wide default hosted tool profile (full|slim).
ROOTLY_MCPCAT_PROJECT_ID optional Optional; hosted deployment project ID for sending MCP telemetry to AgentCat.
  1. Use hosted service: No local installation required, use URL: https://mcp.rootly.com/mcp
  2. Local installation:
  • Install Python 3.12 or higher
  • Install uv package manager
  • Set ROOTLY_API_TOKEN environment variable
  • Run with uvx command: uvx --from rootly-mcp-server rootly-mcp-server
claude_desktop_config.json 
{
  "hosted": {
    "url": "https://mcp.rootly.com/mcp",
    "auth": "OAuth2 or API Token"
  },
  "local": {
    "command": "uvx",
    "args": [
      "--from",
      "rootly-mcp-server",
      "rootly-mcp-server"
    ],
    "env": {
      "ROOTLY_API_TOKEN": "Your Rootly API token"
    }
  }
}

Shown for Claude Desktop. Other clients may use a different file or key (VS Code uses "servers") — the configurator below converts it.

Generate and validate client config

Check that it works

After connecting, check that tools like get_server_version and list_incidents appear in your client's tool list, or run a list_incidents query — a successful response proves the connection. Alternatively, run `uv run python -m rootly_mcp_server --list-tools` to see the effective tool names.

Troubleshooting

  1. Check if API token is correctly set
  2. Verify network connection to Rootly service
  3. Confirm MCP client configuration is correct
  4. Check Rootly service status
  5. Review logs for detailed error information

Things to try

Once connected, you can ask your AI assistant things like:

  • List all unresolved incidents in Rootly
  • Find historical incidents similar to the recent database outage
  • Who is on call right now across all teams?
  • Create an action item for this incident

Tools 100

list_incidents read-only
List all incidents
get_incident read-only
Get detailed information about a specific incident
create_incident writes
Create a new incident
update_incident writes
Update an existing incident
search_incidents read-only
Search for incidents based on criteria
find_related_incidents read-only
Find historically similar incidents using similarity analysis
suggest_solutions read-only
Recommend solutions based on past successful resolutions
create_incident_action_item writes
Create an incident action item
Show 92 more tools
list_incident_action_items read-only
List action items for an incident
update_incident_form_field_selection writes
Update incident form field selections
list_teams read-only
List all teams
get_current_user read-only
Get current user information
list_services read-only
List all services
list_severities read-only
List all severity levels
get_alert read-only
Get detailed information about a specific alert
list_alerts read-only
List all alerts
get_alert_by_short_id read-only
Get an alert by its short ID
list_escalation_policies read-only
List all escalation policies
get_escalation_policy read-only
Get detailed information about a specific escalation policy
list_on_call_roles read-only
List all on-call roles
list_schedules read-only
List all schedules
get_schedule_shifts read-only
Get schedule shifts
get_oncall_handoff_summary read-only
Get on-call handoff summary
get_shift_incidents read-only
Get incidents during a shift
list_endpoints read-only
List all endpoints
check_oncall_health_risk read-only
Check on-call health risk
check_responder_availability read-only
Check responder availability
collect_incidents read-only
Collect incident data
create_override_recommendation read-only
Create override recommendation
get_server_version read-only
Get server version information
list_workflow_runs read-only
create_incident_form_field_selection writes
create_workflow_task writes
get_alert_event read-only
get_alert_group read-only
get_alert_routing_rule read-only
get_alert_source read-only
get_alert_urgency read-only
get_catalog read-only
get_catalog_entity read-only
get_cause read-only
get_custom_form read-only
get_environment read-only
get_escalation_level read-only
get_escalation_path read-only
get_form_field read-only
get_form_field_option read-only
get_functionality read-only
get_functionality_incidents_chart read-only
get_functionality_uptime_chart read-only
get_incident_action_items read-only
get_incident_type read-only
get_on_call_role read-only
get_on_call_shadow read-only
get_override_shift read-only
get_schedule read-only
get_schedule_rotation read-only
get_service read-only
get_service_incidents_chart read-only
get_service_uptime_chart read-only
get_severity read-only
get_status_page read-only
get_status_page_template read-only
get_team read-only
get_team_incidents_chart read-only
get_user read-only
get_workflow read-only
get_workflow_form_field_condition read-only
get_workflow_group read-only
get_workflow_task read-only
list_alert_events read-only
list_alert_groups read-only
list_alert_routing_rules read-only
list_alert_sources read-only
list_alert_urgencies read-only
list_all_incident_action_items read-only
list_catalog_entities read-only
list_catalogs read-only
list_causes read-only
list_custom_forms read-only
list_environments read-only
list_escalation_levels read-only
list_escalation_levels_paths read-only
list_escalation_paths read-only
list_form_field_options read-only
list_form_fields read-only
list_functionalities read-only
list_incident_alerts read-only
list_incident_form_field_selections read-only
list_incident_types read-only
list_on_call_shadows read-only
list_override_shifts read-only
list_schedule_rotation_active_days read-only
list_schedule_rotation_users read-only
list_schedule_rotations read-only
list_shifts read-only
list_status_page_templates read-only
list_status_pages read-only
list_users read-only
list_workflow_form_field_conditions read-only

Use cases

Incident response and management
On-call scheduling and coverage
Alert management and monitoring
Incident analysis and trending
Post-incident analysis and reviews
Service health monitoring
Team collaboration and communication

Supported clients

Claude Desktop
Claude Code
Cursor
Windsurf
Gemini CLI
Codex

Listed from the project's documentation, not tested by this site.

Overview

Rootly MCP Server provides incident management, on-call scheduling, and intelligent analysis powered by Rootly API for Cursor, Windsurf, Claude, and other MCP clients.

Similar servers

Context7 80 · B

Upstash's official server providing up-to-date third-party library docs for AI coding assistants

★ 63.3k · Tools 2 Compare with this →
Compare with other MCP servers

Source revision 8b3ac4cc3a47 Data synced 2026-10-19 Read the FMRS scoring method