Hello! This page is a semi-organized mess of notes while learning PostgreSQL.

Below are PostgreSQL-tagged blog posts:

PostgreSQL (10)


I keep queries in a GitHub repository: pg_scripts

Query: Approximate Count

Since a COUNT(*) query can be slow, try an approximate count:

SELECT reltuples::numeric AS estimate
FROM pg_class WHERE relname = 'table_name';

Query: Get Table Stats

FROM pg_stats
WHERE tablename = 'table';

Consider cardinality for columns, and selectivity for indexes and queries, when designing indexes.

Cancel or Kill by Process ID

Get a PID with SELECT * FROM pg_stat_activity;

Try to cancel the query first, otherwise terminate the backend:

SELECT pg_cancel_backend(pid); 
SELECT pg_terminate_backend(pid);

Tuning Autovacuum

PostgreSQL runs a scheduler Autovacuum process when PostgreSQL starts up. This process looks at configurable thresholds for all tables and determines whether a VACUUM worker should run, per table.

Thresholds can be configured per table. A good starting place for tables that have a large amount of UPDATE and DELETE queries, is to perform that per-table tuning.

The goal is to make VACUUM run more regularly and for a longer period of time, to stay on top of the accumulation of bloat from dead tuples, so that operations are reliable and predictable.

In Routine Vacuuming, the two options are listed:

The scale factor defaults to 20% (0.20). For large tables with a high amount of updates and deletes, we lowered the value to 1% (0.01). With the lowered threshold, vacuum will run more frequently in proportion to how much it’s needed.

Set the value for a table:

ALTER TABLE bigtable SET (autovacuum_vacuum_scale_factor = 0.1);

Can be reset:

ALTER TABLE bigtable RESET (autovacuum_vacuum_threshold);

Autovacuum Tuning

Specialized Index Types

The most common type is B-Tree. Specialized index types are:

  • Multicolumn
  • Covering (Multicolumn or newer INCLUDES style)
  • Partial
  • GIN
  • GiST
  • BRIN
  • Expression
  • Unique
  • Hash

Removing Unused Indexes

Ensure these are set to on

SHOW track_activities;
SHOW track_counts;

Cybertec blog post with SQL query to discover unused indexes: Get Rid of Your Unused Indexes!

Remove Duplicate and Overlapping Indexes


Remove Seldom Used Indexes on High Write Tables

New Finding Unused Indexes Query

This is a great guideline.

As a general rule, if you’re not using an index twice as often as it’s written to, you should probably drop it.

In our system on our highest write table we had 10 total indexes defined and 6 were classified as Low Scans, High Writes. These indexes may not be worth keeping.

Partial Indexes

How Partial Indexes Affect UPDATE Performance in PostgreSQL

Timeout Tuning

  • statement_timeout: The maximum time a statement is allowed to run before being canceled
  • lock_timeout

Connections Management

A connection forks the OS process (creates a new process) and is thus expensive.

Use a connection pooler to reduce overhead from connection establishment

Connection issues could benefit from changing:

  • connect_timeout
  • read_timeout
  • checkout_timeout (Rails, default 5s): maximum time Rails will spend trying to check out a connection from the pool before raising an error. checkout_timeout API documentation
  • statement_timeout. In Rails/Active Record, set in config/database.yml under a variables section with a value in milliseconds. This becomes a session variable which is set like this:

SET statement_timeout = 5000 (in milliseconds) and be displayed like this: SHOW statement_timeout

    statement_timeout: 5000

For Rails with Puma and Sidekiq, carefully manage the connection pool size and total connections.

The Ruby on Rails database connection pool. We also use a proxy in between the application and PG.


Install PgBouncer on macOS with brew install pgbouncer. Create the .ini config file as the article mentions, point it to a database, accept connections, and track the connection count.

Heap Only Tuple (HOT) Updates

HOT (“heap only tuple”) updates, are updates to tuples not referenced from outside the table block.

HOT updates in PostgreSQL for better performance

2 requirements:

  • there must be enough space in the block containing the updated row
  • there is no index defined on any column whose value is modified (big one)


What is fillfactor and how does it affect PostgreSQL performance?

  • Percentage between 10 and 100, default is 100 (“fully packed”)
  • Reducing it leaves room for “HOT” updates when they’re possible. Set to 90 to leave 10% space available for HOT updates.
  • “good starting value for it is 70 or 80” Deep Dive
  • For tables with heavy updates a smaller fillfactor may yield better write performance
  • Set per table or per index (B-Tree defaults to 90 fillfactor)
  • Trade-off: “Faster UPDATE vs Slower Sequential Scan and wasted space (partially filled blocks)” from Fillfactor Deep Dive
  • No index defined on any column whose value is modified

Limitations: Requires a VACUUM FULL after modifying (or pg_repack)

ALTER TABLE foo SET ( fillfactor = 90 );

Or use pg_repack

pg_repack --no-order --table foo

Installing pg_repack on EC2 for RDS

Note: use -k, --no-superuser-check

Locks Management

Lock Monitoring

  • log_lock_waits
  • deadlock_timeout

“Then slow lock acquisition will appear in the database logs for later analysis.”

Lock Types

Exclusive locks, and shared locks.

AccessExclusiveLock - Locks the table, queries are not allowed.

Table locks and row locks.


Tools: Query Planning


This article 5 Things I wish my grandfather told me about ActiveRecord and PostgreSQL has a nice translation of EXPLAIN ANLAYZE output written more in plain English.


YouTube demonstration video

Format EXPLAIN output with JSON, and specify additional options.


Using pgbench

Repeatable method of determining a transactions per second (TPS) rate.

Useful for determining impact of parameter tuning. Configurable with custom SQL queries.

Could be used to test the impact of increasing concurrent connections.

  • Initialize database example with scaling option of 50 times the default size:
pgbench -i -s 50 example`
  • Benchmark with 10 clients, 2 worker threads, and 10,000 transactions per client:
pgbench -c 10 -j 2 -t 10000 example`

I created PR #5388 adding pgbench to tldr!


PGTune is a website that tries to suggest values for PG parameters.



PgHero brings operational concerns into a web dashboard. Built as a Rails engine.

We’re running it in production and saw immediate value in identifying unused and duplicate indexes to remove.

Fix Typo PR #384





Perl script to analyze a database. Has some insights like the shared buffer hit rate, index analysis, configuration advice, and extension recommendations.







brew install pgcli

An alternative to psql with syntax highlighting, autocomplete and more.

Write Ahead Log (WAL) Tuning

Can cause a significant I/O load

  • checkpoint_timeout - in seconds, default checkpointing every 5 minutes
  • max_wal_size - if max wal size is about to be exceeded, default 1 GB

Reducing the value causes more frequent checkpoint operations.

checkpoint_warning parameter checkpoint_completion_target

General Recommendation (not mine)

“On a system that’s very close to maximum I/O throughput during normal operation, you might want to increase checkpoint_completion_target to reduce the I/O load from checkpoints.”


  • commit_delay (0 by default)
  • wal_sync_method
  • wal_debug

Extensions and Modules

Foreign Data Wrapper (FDW)

Native Foreign Data Wrapper (FDW) functionality in PostgreSQL allows connecting to remote sources

The table structure may be specified when establishing the foreign table

We were able to avoid the need for any intermediary data dump files.

We used a temp schema to isolate temporary tables away from the main schema (public).

The process is:

  1. Create a server
  2. Create a user mapping
  3. Create a foreign table (optionally importing the schema)

Let’s say we had 2 services, one for managing inventory items for sale, and one for managing authentication.

Connect as a superuser

create EXTENSION postgres_fdw;


CREATE SERVER temp_authentication;
OPTIONS (host 'authentication-db-host', dbname 'authentication-db-name', port '5432'); -- set the host, name and port

SERVER temp_authentication
OPTIONS (user 'authentication-db-user', password 'authentication-db-password'); -- map the local postgres user to a user on the remote DB

    FROM SERVER temp_authentication INTO temp; -- this will make a table called temp.customers


Generate universally unique identifiers (UUIDs) in PostgreSQL.

Documentation link


Tracks execution stats for all statements. Stats made available from view. Requires reboot (static param) on RDS on PG 10 although pg_stat_statements is available by default in shared_preload_libraries in PG 12.




The pgstattuple module provides various functions to obtain tuple-level statistics.



Case insensitive column type



Available on PG 12.5+ on RDS, pg_cron is an extension that can be useful to schedule maintenance tasks, like manual vacuum jobs.

See: Scheduling maintenance with the PostgreSQL pg_cron extension


pg_timetable: Advanced scheduling for PostgreSQL



Replacement for pg_repack, automated, without needing to run a CLI tool.


PG 10 auto_explain

Adds explain plans to the PostgreSQL log.

Percona pg_stat_monitor

pg_stat_monitor: A cool extension for better monitoring using PMM - Percona Live Online 2020

pganalyze Index Advisor

A better way to index your PostgreSQL database: pganalyze Index Advisor


brew install pgbadger



How does bloat (table bloat, index bloat) affect performance?

  • “When a table is bloated, PostgreSQL’s ANALYZE tool calculates poor or inaccurate information that the query planner uses.”. Example of 7:1 bloated/active tuples ratio causing query planner to skip.
  • Queries on tables with high bloat will require additional IO, navigating through more pages of data.
  • Bloated indexes, such as indexes that reference tuples that have been vacuumed, add IO. Rebuild the index REINDEX ... CONCURRENTLY
  • Index only scans slow down with outdated statistics. Autovacuum updates table statistics. Minimize table bloat to improve performance of index only scans. PG Routing vacuuming docs.
  • Cybertec: Detecting Table Bloat
  • Dealing with significant PostgreSQL database bloat — what are your options?


This is also a really cool Version Upgrade Comparison Tool: 10 to 12

PG 11

October 2018

  • Improves parallel query performance and parallelism of B-tree index creation
  • Adds partitioning by hash key
  • Significant partitioning improvements
  • Adds “covering” indexes with INCLUDE keyword. Adds “payload columns” to indexes. Docs: Index only scans and Covering indexes

PG 12

Release announcement. Released October 2019.

  • Partitioning performance improvements
  • REINDEX gains support for CONCURRENTLY. This is very helpful for online index rebuilds, to remove “bloat” from index entries.

PG 13

Released September 2020

  • Parallel vacuum

PG 14

PG 15

  • SQL MERGE command for Upserts!

PG 16

Released September 2023


Amazon RDS hosts PostgreSQL (with customizations). RDS is a regular single-writer primary instance model for PostgreSQL.

Aurora PostgreSQL

  • Separates storage and compute
  • “Fast clones”
  • Regional offering (single region) and Global offering (multi-region)
  • Can create multi-region replication natively, or without the need for VPC Peering

AWS RDS Parameter Groups

Working with RDS Parameter Groups

  • Try out parameter changes on a test database prior to making the change. Create a snapshot before making the change.
  • Parameter groups can be restored to their defaults (or they can be copied to create an experimental group). Groups can be compared with each other to determine differences.
  • Parameter values can process a formula. RDS provides some formulas that use the instance class CPU or memory available to calculate a value.

Database Constraints

Blog: A Look at PostgreSQL Foreign Key Constraints


Native Replication

  • Physical Replication
  • Logical Replication using PUBLICATION and SUBSCRIPTION objects.
  • PostgreSQL 16 gained bi-directional logical replication for the same table, on two separate instances!

PostgreSQL Logical Replication

Crunchydata Logical Replication in PostgreSQL

  • Create a PUBLICATION, counterpart SUBSCRIPTION.
  • All operations like INSERT and UPDATE are enabled by default, fewer can be configured
  • Logical replication available since PG 10.
  • max_replication_slots should be set higher than number of replicas
  • A role must exist for replication
  • Replication slot is a replication object that keeps track of where the subscriber is in the WAL stream
  • Unlike normal replication, writes are still possible to the subscriber. Conflicts can occur if data is written that would conflict with logical replication.

Declarative Partitioning

Partition Pruning

Default is on or SET enable_partition_pruning = off; to turn it off.


Uncategorized Content

  • Use NULLs instead of default values when possible, cheaper to store and query. Source: Timescale DB blog

Stored Procedures

Stored Procedures are User Defined Functions (UDF).

Using PL/pgSQL, functions can be added to the database directly. Procedures and functions can be written in other languages.

Stored procedures

To manage these functions in a Ruby app, use the fx gem (versioned database functions)!

PostgreSQL Monitoring

  • pg_top On Mac OS
  • brew install pg_top and run it pg_top

Uncategorized Resources

This is an amazing article full of nuggets.

  • The idea of an “Application DBA”
  • Things I liked: Using an intermediate table for de-duplication. Column structure is elegant, clearly broken out destination ID and nested duplicate IDs.
  • Working with arrays
    • ANY() for an array of items to compare against
    • array_remove(anyarray, anyelement) to build an array but remove an element
    • array_agg(expression) to build up list of IDs and unnest(anyarray) to expand it
  • Avoidance of indexes for low selectivity, and value of partial indexes in those cases (activated 90% v. unactivated users 10%)
  • Tip on confirming index usage by removing index in a transaction with BEGIN and rolling it back with ROLLBACK.

  • Generalists/specialists: Application DBA and Performance Analyst
  • PostgreSQL Connection Pooling: Part 1 – Pros & Cons

PostgreSQL Presentations

PostgreSQL Tuning


shared_buffers. RDS default is around 25% of system memory. Recommendations say up to 40% of system memory could be allocated, at which point there may be diminishing returns beyond that.

The unit is 8kb chunks, and requires some math to change the value for. Here is a formula:


Parameter Unit Default RDS Tuned Link
shared_buffers 8kb 25% mem    
autovacuum_cost_delay ms 20 2  
autovacuum_vaccum_cost_limit   200 2000 Docs
effective_cache_size 8kb      
work_mem MB 4 250  
min_wal_size MB 80 4000 High write log blog
max_wal_size MB 4000 16000  
max_worker_processes   8 1x/cpu  
max_parallel_workers   8 1x/cpu  
max_parallel_workers_per_gather   2 4  

PostgreSQL Backups


Identity Columns

  • Identity columns are recommended for primary keys, over using Sequences (with Serial or BigSerial)

Scaling Web Applications

Full Text Search (FTS)

Engineering blog posts

Big companies use PostgreSQL and Ruby on Rails.

Sharding and Scaling

Operational Excellence