PostgreSQL Field Notes

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.

PostgreSQL Topic Hubs

PostgreSQL Guide (Pillar)

pgvector Monitoring Query Optimizer Sizing Slow Query Monitoring Query Monitoring

PostgreSQL Tools

HNSW Tuning Lab Plan Autopsy Index Rollout

Tool Comparisons

Compare PostgreSQL monitoring tools

Browse by topic

pgvector and RAG Indexes Slow Queries Locks and Transactions Backups and Durability Replication and WAL Connections and Pooling CPU, Memory, and I/O Vacuum and Bloat Cloud Providers

February 12, 20266 min read

Azure Database for Postgres Flexible Server: The Settings Worth Tuning

Azure Flex's defaults are conservative. The Server parameters blade is where most of the meaningful tuning happens. Here are the parameters I always touch.

February 11, 20267 min read

Running Postgres on Kubernetes: Lessons From the Operators That Almost Worked

Postgres on Kubernetes is feasible now in a way it was not five years ago. The operators are mature, the storage is good enough, and the failure modes are tractable. Here is what to know.

February 10, 20265 min read

N+1 Queries in Postgres: The Bug ORMs Make Easy and the Fix That Sticks

N+1 is the most common ORM-induced performance bug. The query count tells the story; the application code makes it impossible to spot at review time.

February 9, 20266 min read

Prepared Statements in Postgres: When the Generic Plan Bites

Prepared statements skip the planning step on repeated execution. Sometimes that is a 5x speedup. Sometimes it is a 50x slowdown. Knowing the difference matters.

February 8, 20265 min read

Transactions Around HTTP Calls in Postgres: Don't

Wrapping a transaction around an external API call sounds careful and is actually one of the worst patterns in production database code.

February 7, 202613 min read

PostgreSQL Backup Strategies: Restores, WAL, and Test Drills

Backups only matter if restore is rehearsed. The real work is choosing RPO/RTO, preserving WAL, testing restores, and knowing what data you can afford to lose.

February 6, 20266 min read

Application-Side Retry Logic for Postgres: What to Retry, What Not To

Most application-side retry logic is wrong. It either retries everything (and corrupts data) or nothing (and surfaces transient failures to users). Here is the right framework.

February 5, 20266 min read

Batch Jobs Against Postgres: Production Traffic With Worse Manners

A batch job that worked fine in dev can saturate production. The fixes are not exotic — chunk size, lock duration, retry budget — but most teams skip them.

February 4, 20266 min read

Postgres as a Job Queue: How SKIP LOCKED Changed the Game

Using a Postgres table as a job queue used to be a recipe for contention. SKIP LOCKED makes it tractable. Here is the pattern that actually scales.

February 3, 20268 min read

Advisory Locks in Postgres: A Simple Tool That Deserves Respect

Advisory locks are useful when the database cannot infer the thing you are protecting. They are also dangerous when session scope, pooling, and lock ordering are treated casually.

February 2, 20265 min read

Postgres Arrays: When They Help and When They Hide a Modeling Problem

Arrays are a Postgres feature that can simplify code and a feature that can lock you into bad design. Here is the rule I use to decide when to reach for them.

February 1, 20266 min read

JSONB vs Relational Columns: The Tradeoff Most Schemas Get Wrong

JSONB is great for things that genuinely vary per row. It is wrong for things that are stable. Most teams use it for both, and pay for it in queries that should have been simple.