Notes for the problems that show up after launch: bad plans, awkward migrations, index debt, vacuum pressure, replica lag, and the small decisions that keep PostgreSQL, MySQL, MariaDB, and SQL Server easier to operate.
Find the queries dragging your Postgres down, read their plans, and fix them in order of impact.
Read the guide →Which indexes to add, which to drop, and how to spot redundant index debt before it costs you writes.
Read the guide →How autovacuum works, when it falls behind, and how to keep bloat from eating your storage and your buffer cache.
Read the guide →Sizing pools, PgBouncer modes, and how to stop idle-in-transaction sessions from exhausting max_connections.
Read the guide →Track lag, WAL throughput, and replication slots before they turn into stale reads or a full disk.
Read the guide →
Backups, restores, PITR, archive gaps, and recovery drills for PostgreSQL.
AWS RDS, Azure Database for PostgreSQL, Google Cloud SQL, AlloyDB, and managed PostgreSQL monitoring notes.
Connection storms, pool saturation, PgBouncer behavior, and max_connections pressure.
CPU, memory, disk, temp files, checkpoints, buffers, and workload pressure.
Index design, index debt, reindexing, constraint indexes, and evidence-driven DDL.
Lock waits, deadlocks, blocking chains, idle transactions, and transaction hygiene.
Galera clustering, InnoDB and Aria internals, replication, optimizer behavior, and MariaDB operations notes.
InnoDB internals, replication, locking, query performance, and MySQL operations notes.
HNSW, IVFFlat, recall, embedding search, and production RAG performance notes.
Replica lag, WAL growth, failover readiness, hot standby behavior, and replication slots.
Query plans, EXPLAIN ANALYZE, planner regressions, pagination, joins, and statistics.
Wait statistics, tempdb, Query Store, Always On availability groups, and SQL Server operations notes.
Autovacuum, dead tuples, bloat, wraparound, MVCC, freeze age, and visibility.
Finance needed last night's nginx logs joined against orders, and someone remembered the CONNECT engine could read a CSV as a table. It worked beautifully for three weeks — until logrotate moved the file mid-query and every report died with…
The replica stopped with Error 1032 at 21:14, and the on-call runbook said SET GLOBAL sql_slave_skip_counter=1. Six skips later the replica was serving rows that didn't exist on the primary. What these errors actually mean and how to fix the…
Our 5.7 box showed a 38% query cache hit rate and still served most reads from disk — the cache was inflating the numbers while a single mutex capped throughput. What 8.0 removed, how to find dependent queries, and what…
A mysqldump restore died at 61% with ERROR 1153 because one BLOB row exceeded the client default — and the dump itself had been 'successful'. The real defaults per tool, the replication variant, and how to size it without guessing…
A nightly DELETE of 9M rows stalled commits for 40 seconds while every other transaction queued behind it — and lsof showed a 3.1GB deleted-but-open file in /tmp. How the binlog cache works, what the spill costs at commit, and…
We encrypted forty tables for a compliance deadline and nearly lost them to a keyring file that sat in the datadir, unbacked-up. How InnoDB's two-layer keys work, what rotation really rewrites, and the operational rules that keep TDE boring.
The table created fine, then INSERTs started failing with ERROR 1118 only for rows with long JSON payloads. Why InnoDB enforces half a page per row, how DYNAMIC stores columns off-page, the two different 1118 errors, and fixes that survive…
A failover rehearsal exposed 214 rows on a replica that existed nowhere else, and replication had been 'healthy' the whole time. How drift happens under green dashboards, how pt-table-checksum proves it, and how to repair without a rebuild.
I killed a runaway UPDATE at 02:10 and the server kept burning CPU until 04:40 rolling it back. What KILL actually sets, how to measure rollback progress, why restarting makes it worse, and the query shapes that never get killed…
A monitoring agent asking information_schema.TABLES for table sizes every 30 seconds was causing the latency spikes it was supposed to detect. Why metadata reads cost real work, the two knobs that tame them, and how to find the offenders.