@sentry/mcp-server

Exposes Sentry's error event API through the Model Context Protocol, allowing LLM agents to query and filter error events by project, environment, release, and time range. Implements MCP resource and tool patterns to translate natural language queries into Sentry API calls, returning structured event data with stack traces, breadcrumbs, and user context for analysis and debugging.

Exposes Sentry issue operations (resolve, ignore, assign, update status) as MCP tools callable by LLM agents. Translates agent actions into Sentry API mutations, enabling automated issue triage, assignment, and status updates based on error analysis. Implements idempotent operations with proper error handling for concurrent modifications.

Provides MCP resources that expose Sentry project metadata, team structure, and organization configuration to LLM agents. Implements resource listing and hierarchical navigation (organization → projects → teams) to give agents awareness of available projects, their settings, and team ownership. Caches metadata to reduce API calls during multi-step workflows.

Translates natural language queries from LLM agents into Sentry's filter syntax and API parameters. Implements semantic understanding of error attributes (environment, release, user, error type) and maps them to Sentry's query language. Handles ambiguous queries by providing clarification options or reasonable defaults.

Analyzes error events from multiple Sentry projects to identify common patterns, root causes, and correlated failures. Implements statistical analysis of error fingerprints, stack traces, and metadata to group related errors and detect systemic issues. Provides insights into error trends and impact across the organization.

Integrates Sentry release data with deployment information to correlate errors with specific code changes and deployments. Implements release-aware error filtering and analysis, allowing agents to understand which errors were introduced in which release and which deployments are affected. Supports linking errors to commits and pull requests.

Exposes Sentry's performance monitoring data (transactions, spans, latency metrics) through MCP tools, allowing agents to analyze application performance alongside error data. Implements transaction filtering, span analysis, and performance trend detection. Correlates performance degradation with errors to identify performance-related root causes.

Exposes Sentry's user impact data and session replay information through MCP, allowing agents to understand which users are affected by errors and access session context for debugging. Implements user filtering, session correlation, and impact metrics. Provides links to session replays for visual debugging.

Exposes Sentry's alert rules and notification settings through MCP, allowing agents to configure alerts, manage notification channels, and adjust alerting thresholds. Implements alert rule creation, modification, and testing. Supports multiple notification channels (email, Slack, PagerDuty, webhooks).

Integrates Sentry error data with external incident management platforms (PagerDuty, Opsgenie, etc.) through MCP, enabling agents to create incidents, update severity, and manage escalations based on error analysis. Implements bidirectional sync of incident status and error resolution.