MCP SERVERDiagnose a Postgres or MySQL database — unused and missing indexes, table bloat, slow queries, idle-in-transaction connections, and VACUUM status — as findings rather than statistics.
LABSCO SUMMARY
A session sitting idle inside a transaction holds its locks, prevents VACUUM from reclaiming dead tuples, and looks completely normal in a connection count. It is a specific piece of operational knowledge encoded as a tool, and the same instinct runs through the rest: unused indexes and missing indexes as separate analyses because they are separate decisions, bloat measured differently per engine because the engines fragment differently. The pairing of analyze_vacuum with bloat and idle-in-transaction is the whole PostgreSQL maintenance story in three calls. Read-only, which is the right scope for a diagnostic.
A database diagnostic layer covering the problems that accumulate silently: index usage, bloat, query plans, slow-query statistics, connection state, foreign-key structure and PostgreSQL maintenance.
DATABASE_URL. Everything here reads — the tools inspect and analyse, and none of them alter the schema or the data.
One command plus credentials — npx -y mcp-db-analyzer, then supply credentials
USE CASES
Read the slowest queries from the server's own statistics, then explain the worst one with its cost estimates.
Find the unused ones wasting space and, separately, the tables scanning sequentially with no index usage at all — two different decisions.
Check table bloat by dead tuple ratio or InnoDB fragmentation, and read the VACUUM status that explains why it accumulated.
Look for idle-in-transaction sessions specifically — they hold locks and block maintenance while looking like ordinary connections.
List the tables with row counts and sizes, then build the foreign-key dependency graph to see how they hang together.
Counted once with only startup setup. Other versions, modes or optional credentials may expose a different list.
analyze_connectionsAnalyze active database connections. Detects idle-in-transaction sessions and lock contention between sessions (PostgreSQL), long-running queries flagged at >30 seconds, and connection pool utilization. Idle-in-transaction and lock contention detection are not available on MySQL — use this tool's output to investigate PostgreSQL-specific blocking scenarios. Not available for SQLite.
analyze_indexesAnalyze index usage statistics to find unused indexes wasting space and missing indexes causing slow sequential scans. Also detects unindexed foreign keys.
analyze_slow_queriesFind the slowest queries using pg_stat_statements (PostgreSQL) or performance_schema (MySQL). Shows execution times, call counts, and optimization recommendations. PostgreSQL requires the pg_stat_statements extension to be installed and listed in shared_preload_libraries — the tool returns setup instructions if the extension is missing. Not available for SQLite.
analyze_table_bloatAnalyze table bloat by checking dead tuple ratios (PostgreSQL) or InnoDB fragmentation (MySQL), vacuum history, and table sizes.
analyze_table_relationshipsAnalyze foreign key relationships between tables. Builds a dependency graph showing entity connectivity, orphan tables (no FKs), cascading delete chains (shown at full depth), hub entities (tables with 5+ FK connections), and circular FK dependencies. Useful for understanding schema design, planning migrations, and impact analysis.
analyze_vacuumAnalyze PostgreSQL VACUUM maintenance status. Checks dead tuple ratios, vacuum staleness, autovacuum configuration, and identifies tables needing manual VACUUM. PostgreSQL only.
explain_queryRun EXPLAIN on a SQL query and return a formatted plan with cost estimates, node types, and optimization warnings. Optionally runs EXPLAIN ANALYZE for actual execution statistics (read-only queries only).
inspect_schemaList all tables in a schema with row counts and sizes, or inspect a specific table's columns, types, constraints, and foreign keys.
suggest_missing_indexesFind tables with high sequential scan counts and zero index usage, cross-referenced with unused indexes wasting space. Provides actionable CREATE INDEX and DROP INDEX recommendations.
BEFORE YOU INSTALL
9 tools when we connected.
All 3 documented setups run the server on your machine.
CONTEXT COST
Rated by the size of the 9 tool definitions that ride along on every turn, not by how many there are
INSTALLATION
npx -y mcp-db-analyzer{
"mcpServers": {
"db-analyzer": {
"command": "npx",
"args": [
"-y",
"mcp-db-analyzer"
],
"env": {
"DATABASE_URL": "postgresql://<YOUR_DB_USER>:<YOUR_DB_PASSWORD>@localhost:5432/<YOUR_DB_NAME>"
}
}
}
}macOS ~/Library/Application Support/Claude/claude_desktop_config.json · Windows %APPDATA%\Claude\claude_desktop_config.json
{
"mcpServers": {
"db-analyzer": {
"command": "npx",
"args": [
"-y",
"mcp-db-analyzer",
"--driver",
"mysql"
],
"env": {
"DATABASE_URL": "mysql://<YOUR_DB_USER>:<YOUR_DB_PASSWORD>@localhost:3306/<YOUR_DB_NAME>"
}
}
}
}{
"mcpServers": {
"db-analyzer": {
"command": "npx",
"args": [
"-y",
"mcp-db-analyzer",
"--driver",
"sqlite"
],
"env": {
"DATABASE_URL": "<YOUR_DATABASE_URL>"
}
}
}
}MCP servers
ahmetbarut
NewToolAI
TheRaLabs
crystaldba