Description

Add MCP-PostgreSQL-Ops: a professional, read-only MCP server for PostgreSQL operations, monitoring, and performance analysis. It surfaces server info, active connections, configuration search, capacity metrics, and pg_stat_statements / pg_stat_monitor–based insights via safe, predefined tools.

Server Details

• Server: postgresql-ops (Python, FastMCP; supports stdio and streamable-http)
• Changes to: New server entry with tools for server status, structure exploration, performance monitoring, and capacity management.

Motivation and Context

Many PostgreSQL MCP servers focus on general SQL access. MCP-PostgreSQL-Ops is purpose-built for operations and observability in production: it provides slow-query analysis, index-usage insights, config discovery, and size/capacity views—all read-only to minimize risk. This fills a gap for teams that want actionable diagnostics without enabling write access from LLM clients.

How Has This Been Tested?

Tested locally with:
Claude Desktop using the provided mcpServers configuration; verified tool calls and end-to-end responses.
MCP Inspector via the included script for interactive tool validation.

Scenarios exercised:
get_server_info, get_active_connections, get_postgresql_config
get_pg_stat_statements_top_queries, get_pg_stat_monitor_recent_queries, get_index_usage_stats
get_database_size_info, get_table_size_info, get_vacuum_analyze_stats

Both stdio and streamable-http modes were run as documented.

Breaking Changes

None. This is an additive server entry; no existing MCP clients need to change unless they opt in by adding the server to their mcpServers configuration.

Types of changes

• [] Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to change)
• Documentation update

Checklist

• I have read the MCP Protocol Documentation
• My changes follows MCP security best practices
• I have updated the server's README accordingly
• I have tested this with an LLM client
• My code follows the repository's style guidelines
• New and existing tests pass locally
• I have added appropriate error handling
• I have documented all environment variables and configuration options

Additional context

• Key features/tools: server info & sessions, structure exploration (DBs/tables/users), performance (top queries, real-time monitoring, index usage), and capacity (DB/table sizes, VACUUM/ANALYZE stats).
• Security posture: strictly read-only; no direct SQL—only predefined queries; sensitive fields masked.
• Transport modes: stdio by default; streamable-http supported for local testing.
• Env vars: POSTGRES_HOST, POSTGRES_PORT, POSTGRES_USER, POSTGRES_PASSWORD, POSTGRES_DB, plus FASTMCP_* and MCP_LOG_LEVEL.
• Prereqs: PostgreSQL ≥12 (tested with 16); recommended extensions pg_stat_statements (required for some tools) and optional pg_stat_monitor.