· 16 min · PostgreSQL

Managed Postgres, Examined: Azure Database for PostgreSQL Flexible Server

Azure's managed PostgreSQL differs from competitors by putting the standby in the commit path—every write waits for synchronous replication to a second server…

0 comments

· 3 min · PostgreSQL

All Your GUCs in a Row: cpu_index_tuple_cost, cpu_operator_cost, and cpu_tuple_cost

cpu_tuple_cost, cpu_index_tuple_cost, and cpu_operator_cost are three of the constants the planner uses to price a query, and the single most useful thing to kn

0 comments

· 5 min · PostgreSQL

SQL/PGQ in PostgreSQL 19: Graph Queries Without the Graph Database

PostgreSQL 19 adds GRAPH_TABLE, letting you query property graphs with Cypher-like pattern matching over your existing relational tables.

0 comments

· 3 min · PostgreSQL

All Your GUCs in a Row: constraint_exclusion

Skip partition scanning with constraint_exclusion, PostgreSQL's old pruning trick.

0 comments

· 2 min · PostgreSQL

All Your GUCs in a Row: config_file

PostgreSQL's `config_file` parameter creates a bootstrap paradox: it tells the server where to find its configuration, but lives on the command line only—never…

0 comments

· 3 min · PostgreSQL

All Your GUCs in a Row: compute_query_id

PostgreSQL 14 unified query-id computation across all subsystems, but defaulting to always-on would tax every backend.

0 comments

· 6 min · PostgreSQL

Open-Source TDE for PostgreSQL: What pg_tde Is, and Whether You Need It

PostgreSQL finally has an open-source Transparent Data Encryption option.

0 comments

· 2 min · PostgreSQL

All Your GUCs in a Row: commit_timestamp_buffers

PostgreSQL 17 made SLRU buffer pools configurable for the first time.

0 comments

· 5 min · PostgreSQL

Twenty Years, Three CVEs, One AI

Three heap buffer overflows in PostgreSQL — including a 20-year-old pgcrypto bug — were found by an AI code analyzer. But.

0 comments

· 3 min · PostgreSQL

All Your GUCs in a Row: commit_delay and commit_siblings

Tune `commit_delay` to batch WAL flushes and trade latency for throughput—but only if `pg_test_fsync` proves sync time is your bottleneck.

0 comments