-- ============================================================================
-- Slow Query Analysis Script
-- Purpose: Identify and analyze slow-running queries (including 110-second ones)
-- Requirements: pg_stat_statements extension must be enabled
-- ============================================================================
\pset pager off

\timing on
\x auto

\echo '========================================='
\echo 'Jellyfin Slow Query Analysis'
\echo 'Generated: 2026-03-05'
\echo '========================================='
\echo ''

-- ============================================================================
-- 1. CHECK EXTENSION STATUS
-- ============================================================================

\echo '1. CHECKING pg_stat_statements EXTENSION'
\echo '-----------------------------------------'

SELECT
    CASE 
        WHEN COUNT(*) > 0 THEN '✅ pg_stat_statements is installed'
        ELSE '❌ pg_stat_statements is NOT installed - Run: CREATE EXTENSION pg_stat_statements;'
    END AS status
FROM pg_extension
WHERE extname = 'pg_stat_statements';

\echo ''

-- ============================================================================
-- 2. TOP 20 SLOWEST QUERIES BY MAX EXECUTION TIME
-- ============================================================================

\echo '2. TOP 20 SLOWEST QUERIES (by max execution time)'
\echo '-----------------------------------------'
\echo 'These are the queries that have taken the longest time to execute'
\echo ''

SELECT
    LEFT(query, 100) AS query_preview,
    calls,
    ROUND(mean_exec_time::numeric, 2) AS avg_ms,
    ROUND(max_exec_time::numeric, 2) AS max_ms,
    ROUND(min_exec_time::numeric, 2) AS min_ms,
    ROUND(stddev_exec_time::numeric, 2) AS stddev_ms,
    ROUND(total_exec_time::numeric, 2) AS total_ms,
    ROUND((100.0 * total_exec_time / SUM(total_exec_time) OVER ())::numeric, 2) AS pct_total,
    rows AS rows_returned
FROM pg_stat_statements
WHERE query NOT LIKE '%pg_stat_statements%'
  AND query NOT LIKE '%pg_catalog%'
  AND dbid = (SELECT oid FROM pg_database WHERE datname = current_database())
ORDER BY max_exec_time DESC
LIMIT 20;

\echo ''

-- ============================================================================
-- 3. TOP 20 SLOWEST QUERIES BY AVERAGE EXECUTION TIME
-- ============================================================================

\echo '3. TOP 20 SLOWEST QUERIES (by average execution time)'
\echo '-----------------------------------------'
\echo 'These queries consistently take a long time'
\echo ''

SELECT
    LEFT(query, 100) AS query_preview,
    calls,
    ROUND(mean_exec_time::numeric, 2) AS avg_ms,
    ROUND(max_exec_time::numeric, 2) AS max_ms,
    ROUND(total_exec_time::numeric, 2) AS total_ms,
    ROUND((100.0 * total_exec_time / SUM(total_exec_time) OVER ())::numeric, 2) AS pct_total,
    rows AS rows_returned,
    CASE 
        WHEN calls > 0 THEN ROUND((rows::numeric / calls), 2)
        ELSE 0
    END AS avg_rows_per_call
FROM pg_stat_statements
WHERE query NOT LIKE '%pg_stat_statements%'
  AND query NOT LIKE '%pg_catalog%'
  AND dbid = (SELECT oid FROM pg_database WHERE datname = current_database())
  AND calls > 1  -- Only queries called multiple times
ORDER BY mean_exec_time DESC
LIMIT 20;

\echo ''

-- ============================================================================
-- 4. QUERIES WITH HIGHEST TOTAL TIME (Performance Impact)
-- ============================================================================

\echo '4. QUERIES WITH HIGHEST TOTAL TIME'
\echo '-----------------------------------------'
\echo 'These queries consume the most total database time'
\echo ''

SELECT
    LEFT(query, 100) AS query_preview,
    calls,
    ROUND(mean_exec_time::numeric, 2) AS avg_ms,
    ROUND(max_exec_time::numeric, 2) AS max_ms,
    ROUND(total_exec_time::numeric, 2) AS total_ms,
    ROUND((100.0 * total_exec_time / SUM(total_exec_time) OVER ())::numeric, 2) AS pct_total,
    ROUND((total_exec_time / 1000 / 60)::numeric, 2) AS total_minutes
FROM pg_stat_statements
WHERE query NOT LIKE '%pg_stat_statements%'
  AND query NOT LIKE '%pg_catalog%'
  AND dbid = (SELECT oid FROM pg_database WHERE datname = current_database())
ORDER BY total_exec_time DESC
LIMIT 20;

\echo ''

-- ============================================================================
-- 5. BASEITEMS QUERIES SPECIFICALLY (Main Problem Area)
-- ============================================================================

\echo '5. BASEITEMS TABLE QUERIES ANALYSIS'
\echo '-----------------------------------------'
\echo 'Focusing on BaseItems table (where 110-second queries occur)'
\echo ''

SELECT
    LEFT(query, 150) AS query_preview,
    calls,
    ROUND(mean_exec_time::numeric, 2) AS avg_ms,
    ROUND(max_exec_time::numeric, 2) AS max_ms,
    ROUND(min_exec_time::numeric, 2) AS min_ms,
    ROUND(total_exec_time::numeric, 2) AS total_ms,
    rows AS rows_returned,
    CASE 
        WHEN calls > 0 THEN ROUND((rows::numeric / calls), 2)
        ELSE 0
    END AS avg_rows_per_call,
    CASE
        WHEN max_exec_time > 100000 THEN '🔥 CRITICAL (>100s)'
        WHEN max_exec_time > 10000 THEN '⚠️ SLOW (>10s)'
        WHEN max_exec_time > 1000 THEN '⚡ MODERATE (>1s)'
        ELSE '✅ FAST'
    END AS performance_rating
FROM pg_stat_statements
WHERE query ILIKE '%BaseItems%'
  AND query NOT LIKE '%pg_stat_statements%'
  AND query NOT LIKE '%pg_catalog%'
  AND dbid = (SELECT oid FROM pg_database WHERE datname = current_database())
ORDER BY max_exec_time DESC;

\echo ''

-- ============================================================================
-- 6. PEOPLES AND ITEMVALUES QUERIES
-- ============================================================================

\echo '6. PEOPLES AND ITEMVALUES QUERIES'
\echo '-----------------------------------------'
\echo 'Queries on high-sequential-scan tables'
\echo ''

SELECT
    CASE 
        WHEN query ILIKE '%Peoples%' THEN 'Peoples'
        WHEN query ILIKE '%ItemValues%' THEN 'ItemValues'
        ELSE 'Other'
    END AS table_name,
    LEFT(query, 100) AS query_preview,
    calls,
    ROUND(mean_exec_time::numeric, 2) AS avg_ms,
    ROUND(max_exec_time::numeric, 2) AS max_ms,
    ROUND(total_exec_time::numeric, 2) AS total_ms,
    rows AS rows_returned
FROM pg_stat_statements
WHERE (query ILIKE '%Peoples%' OR query ILIKE '%ItemValues%')
  AND query NOT LIKE '%pg_stat_statements%'
  AND query NOT LIKE '%pg_catalog%'
  AND dbid = (SELECT oid FROM pg_database WHERE datname = current_database())
ORDER BY max_exec_time DESC
LIMIT 20;

\echo ''

-- ============================================================================
-- 7. QUERIES WITH MANY ROWS BUT FAST (Good Index Usage)
-- ============================================================================

\echo '7. EFFICIENT QUERIES (Many rows, fast execution)'
\echo '-----------------------------------------'
\echo 'These queries return many rows but execute quickly (good indexes!)'
\echo ''

SELECT
    LEFT(query, 100) AS query_preview,
    calls,
    ROUND(mean_exec_time::numeric, 2) AS avg_ms,
    rows AS total_rows_returned,
    ROUND((rows::numeric / NULLIF(calls, 0))::numeric, 2) AS avg_rows_per_call,
    ROUND((mean_exec_time / NULLIF((rows::numeric / NULLIF(calls, 0)), 0))::numeric, 4) AS ms_per_row
FROM pg_stat_statements
WHERE rows > 100000  -- Many rows returned
  AND mean_exec_time < 100  -- But fast execution
  AND calls > 10  -- Called frequently
  AND query NOT LIKE '%pg_stat_statements%'
  AND query NOT LIKE '%pg_catalog%'
  AND dbid = (SELECT oid FROM pg_database WHERE datname = current_database())
ORDER BY rows DESC
LIMIT 10;

\echo ''

-- ============================================================================
-- 8. CURRENTLY RUNNING QUERIES (Real-time)
-- ============================================================================

\echo '8. CURRENTLY RUNNING QUERIES'
\echo '-----------------------------------------'

SELECT
    pid,
    usename,
    application_name,
    NOW() - query_start AS duration,
    state,
    wait_event_type,
    wait_event,
    LEFT(query, 100) AS query_preview
FROM pg_stat_activity
WHERE state != 'idle'
  AND pid != pg_backend_pid()
  AND query NOT LIKE '%pg_stat_activity%'
  AND datname = current_database()
ORDER BY query_start ASC;

\echo ''

-- ============================================================================
-- 9. QUERY STATISTICS SUMMARY
-- ============================================================================

\echo '9. OVERALL QUERY STATISTICS SUMMARY'
\echo '-----------------------------------------'

SELECT
    COUNT(*) AS total_unique_queries,
    SUM(calls) AS total_calls,
    ROUND(SUM(total_exec_time)::numeric / 1000 / 60, 2) AS total_exec_minutes,
    ROUND(AVG(mean_exec_time)::numeric, 2) AS avg_query_time_ms,
    ROUND(MAX(max_exec_time)::numeric, 2) AS slowest_query_ever_ms,
    ROUND(MAX(max_exec_time)::numeric / 1000, 2) AS slowest_query_ever_seconds,
    SUM(rows) AS total_rows_returned,
    ROUND(AVG(rows::numeric / NULLIF(calls, 0))::numeric, 2) AS avg_rows_per_query
FROM pg_stat_statements
WHERE query NOT LIKE '%pg_stat_statements%'
  AND query NOT LIKE '%pg_catalog%'
  AND dbid = (SELECT oid FROM pg_database WHERE datname = current_database());

\echo ''

-- ============================================================================
-- 10. RECOMMENDATIONS
-- ============================================================================

\echo ''
\echo '========================================='
\echo 'RECOMMENDATIONS'
\echo '========================================='
\echo ''
\echo 'Based on the analysis above:'
\echo ''
\echo '1. CRITICAL QUERIES (>100s):'
\echo '   - Review queries marked 🔥 CRITICAL'
\echo '   - Check EXPLAIN ANALYZE for these queries'
\echo '   - Consider query rewrite or additional indexes'
\echo ''
\echo '2. SLOW QUERIES (>10s):'
\echo '   - Review queries marked ⚠️ SLOW'
\echo '   - Check if indexes are being used'
\echo '   - Run: EXPLAIN (ANALYZE, BUFFERS) <query>'
\echo ''
\echo '3. HIGH TOTAL TIME QUERIES:'
\echo '   - Even fast queries with many calls add up'
\echo '   - Consider caching frequently called queries'
\echo '   - Optimize connection pooling'
\echo ''
\echo '4. TO RESET STATISTICS:'
\echo '   Run: SELECT pg_stat_statements_reset();'
\echo '   (Useful after making optimization changes)'
\echo ''
\echo '5. TO GET FULL QUERY TEXT:'
\echo '   This script shows only first 100-150 characters'
\echo '   Query full text from pg_stat_statements table directly'
\echo ''
\echo '========================================='