Jira

You can run a Python-based MCP server that exposes Jira issue data stored in Snowflake through a clean, standardized interface. This server lets you list issues, fetch detailed information for a specific issue, and generate project-wide summaries for quick analytics, all while supporting optional Prometheus metrics for monitoring.

How to use

You interact with the Jira MCP Server by connecting your MCP-enabled client to either a local (stdio) instance or a remote HTTP endpoint. Use the list_issues tool to filter and search Jira issues by project, type, status, priority, or text in the summary/description. Use get_issue_details to retrieve comprehensive information for a single issue by its key. Use get_project_summary to obtain overall statistics and status/priority distributions across all projects.

The server connects to Snowflake to read data from JIRA_ISSUE_NON_PII and JIRA_LABEL_RHAI. Tools are designed to work with non-PII data and provide structured outputs that you can feed into AI assistants or analytics workflows.

How to install

Prerequisites You need Python 3.8 or higher and a container/runtime environment (Podman or Docker) to run containerized deployments if you choose that path. You also need access credentials for Snowflake.

Step 1 — Prepare your environment

Step 2 — Set up environment variables for Snowflake access and MCP configuration

Step 3 — Run the MCP server locally

Step 4 — (Optional) Run via container with Podman as shown in the deployment example

{
  "mcpServers": {
    "jira_mcp_snowflake": {
      "command": "python",
      "args": ["src/mcp_server.py"],
      "env": [
        {"name": "SNOWFLAKE_TOKEN", "value": "YOUR_TOKEN"},
        {"name": "SNOWFLAKE_BASE_URL", "value": "https://your-account.snowflakecomputing.com/api/v2"},
        {"name": "SNOWFLAKE_DATABASE", "value": "your_database_name"},
        {"name": "SNOWFLAKE_SCHEMA", "value": "your_schema_name"},
        {"name": "MCP_TRANSPORT", "value": "stdio"},
        {"name": "ENABLE_METRICS", "value": "true"},
        {"name": "METRICS_PORT", "value": "8000"}
      ]
    }
  }
}

Additional configurations and run options

HTTP endpoint configuration is also supported to connect to a remote MCP server or gateway. The remote MCP URL can be used to subscribe and receive data over a server-sent events endpoint when a remote instance is available.

Here is an example for connecting to a remote MCP server via HTTP. This configuration uses the remote URL and includes an authorization header surrogate for token-based access.

{
  "mcpServers": {
    "jira_mcp_snowflake": {
      "type": "http",
      "name": "jira_mcp_snowflake",
      "url": "https://jira-mcp-snowflake.example.com/sse",
      "args": []
    }
  }
}

Monitoring and security considerations

The server can emit Prometheus metrics when enabled. You can monitor tool usage, Snowflake query performance, and active MCP connections. Expose the metrics endpoint on /metrics and health checks on /health.

Security best practices include storing Snowflake tokens in environment variables, rotating tokens regularly, using HTTPS endpoints for remote access, and following the principle of least privilege for Snowflake access.

Available tools

list_issues

Query Jira issues with optional filtering by project, issue type, status, priority, and text search; limit results to control output size.

get_issue_details

Fetch comprehensive information for a specific Jira issue by key, including metadata, timestamps, time tracking, labels, and components.

get_project_summary

Generate statistics across all projects, including total issue counts and per-project status and priority distributions.