Performance dashboard: running queries / index suggestions / slow queries (if extension loaded)

This should be a new tab type with three sub-tabs for each item to inspect.

1. Running queries - just shows queries that are running and lets a user kill them (if they have perms). Useful for killing slow running analytics queries.
2. Index suggestions - databases track which queries are running without indexes and doing full table scans, we're surfacing this information.
3. Slow queries - which queries SUCK and need to be improved.

Some implementation notes:

• A lot of web frameworks (like rails) include a stacktrace as a comment when they run queries, we should surface this for slow queries so the user can see the cause. I'm not sure if all frameworks do this.
• Some of these require plugins, in those cases we should check for the plugin, then block the tab if we detect the plugin is not installed, we can point them to installation instructions.
• Some of this is done by projects like https://github.com/ankane/pghero, so we can learn from them

Availability for major database engines

Feature support by database

Database	Running Queries + Kill	Index Health Advisor	Slow Query Stats
PostgreSQL	✅ pg_stat_activity + pg_cancel_backend/pg_terminate_backend	✅ pg_stat_user_indexes, pg_indexes for duplicates	⚠️ Needs pg_stat_statements extension enabled
MySQL / MariaDB	✅ information_schema.PROCESSLIST + KILL	⚠️ Usage stats need Performance Schema on; no built-in missing-index advisor	⚠️ Needs Performance Schema or slow query log enabled
SQLite	❌ Embedded, single-process — no concept of other connections to view/kill	⚠️ Can list indexes (PRAGMA index_list) but no usage stats to find unused/duplicate	❌ No stats subsystem — no built-in query timing history
Redis	✅ CLIENT LIST / CLIENT KILL	❌ Not a relational/indexed store — doesn't apply	✅ Built-in SLOWLOG
SQL Server	✅ sys.dm_exec_requests/sys.sysprocesses + KILL <spid>	✅ sys.dm_db_index_usage_stats + missing-index DMVs — best native support	⚠️ Needs Query Store enabled (or sys.dm_exec_query_stats)
MongoDB	✅ db.currentOp() / db.killOp()	⚠️ $indexStats gives usage but no missing/duplicate detection	⚠️ Needs profiler enabled (db.setProfilingLevel)

Slow-query setting scope

Database	Setting	Scope	Restart required?	Beekeeper can toggle it?
PostgreSQL	pg_stat_statements	Server-level — must be in shared_preload_libraries	✅ Yes	❌ No — config file + restart is outside Beekeeper's reach; can only detect it's missing and tell the user
MySQL/MariaDB	Performance Schema (performance_schema variable)	Server-level — read-only variable, startup-only	✅ Yes	❌ No
MySQL/MariaDB	(alternative) slow_query_log / long_query_time	Global, dynamic — SET GLOBAL ...	❌ No	⚠️ Yes, but affects all connections server-wide and needs SUPER/SYSTEM_VARIABLES_ADMIN — not something to flip silently
SQL Server	Query Store	Database-level, dynamic — ALTER DATABASE ... SET QUERY_STORE = ON	❌ No	⚠️ Yes, persists for that DB going forward, needs db_owner-ish permission
MongoDB	Profiler (db.setProfilingLevel)	Per-database, dynamic, in-memory — not persisted across mongod restart	❌ No	⚠️ Yes, needs dbAdmin-ish role; in a replica set/sharded cluster has to be set per-node