Notes for the problems that show up after launch: bad plans, awkward migrations, index debt, vacuum pressure, replica lag, and the small decisions that make PostgreSQL easier to operate.
The bigger the shared_buffers and the more connections mapping them, the more 4KB pages cost you in TLB misses and page-table memory. Huge pages are cheap to set up if you size them honestly.
MaxScale's readwritesplit router gives your application one endpoint and spreads reads across replicas. Here is how routing decisions are made, where session state bites, and how failover coordination works.
One database setting ended our reader-writer blocking overnight and introduced two bugs we only caught in a concurrency soak. RCSI mechanics, the honest migration checklist, and the version-store monitoring that keeps tempdb safe.
A week of error 1205s and a proposed retry loop that would fix nothing. How to pull deadlock graphs from system_health, read victim, owner, and waiter nodes, and kill the three patterns behind most deadlocks.
The nightly REBUILD job I inherited burned 90 GB of transaction log and fixed nothing. What fragmentation still means on flash, what REBUILD actually buys over REORGANIZE, and why updating statistics is usually the real win.
The deploy that doubled p95 left its fingerprints in sys.query_store_runtime_stats. How to configure Query Store so the data is there when you need it, compare intervals to find the regressed plan, and force a plan without fooling yourself.
Since-restart wait totals average your quietest night and your worst deploy into one meaningless number. Here is how I diff sys.dm_os_wait_stats over an interval, drop the benign noise list, and branch on what CXPACKET, PAGEIOLATCH, WRITELOG, and LCK_M_* are actually telling you.
A batch job that got slower as we added workers was queueing on three pages in tempdb. How to tell PAGELATCH from PAGEIOLATCH, size the data files, weigh memory-optimized metadata in 2019, and keep the version store from eating the drive.
JIT compilation is supposed to make big PostgreSQL queries faster, but bad cost estimates fire it up on cheap ones. Here is how to spot the regression and switch it off safely.
Galera's multi-primary promise has a footnote: write the same row on two nodes and one transaction dies at COMMIT. How certification conflicts work, what the counters say, and how to design them out.
The default autovacuum cost settings throttle vacuum to a pace modern hardware laughs at. On big write-heavy tables, that means dead tuples win. Here is how to retune.
MariaDB and MySQL forked over a decade ago and the operational differences are now real: incompatible GTID formats, JSON as a LONGTEXT alias, system-versioned tables, a bundled thread pool, and replication gotchas.