The JusDB Database Engineering Blog

How to size your database connection pool correctly — the HikariCP formula, the variables that determine optimal pool size, and the mistakes that cause connection exhaustion or underutilization.

MySQL 8.0 introduced EXPLAIN FORMAT=TREE — a hierarchical, human-readable query plan format that reveals join order, access methods, and cost estimates in a way that the traditional tabular EXPLAIN cannot.

MySQL 8.0 introduced multi-valued indexes — allowing you to index individual elements of a JSON array so queries using MEMBER OF or JSON_OVERLAPS can use an index.

UUIDs as primary keys can destroy database performance. Learn why 50M rows went from 0.5ms to 2.3ms inserts, and how to fix it with hybrid approaches.

Detect query plan regressions in PostgreSQL using auto_explain, pg_hint_plan, and plan baseline management

ProxySQL pools thousands of application connections into a small backend pool, splits reads to replicas automatically, and lets you change routing without touching application code or restarting MySQL.

Identify bloated indexes in PostgreSQL using pgstattuple, REINDEX CONCURRENTLY, and automated monitoring

pg_partman automates PostgreSQL partition lifecycle. This guide covers create_parent(), retention policies, partition pruning, and migrating existing tables without downtime.

Tune PostgreSQL checkpoint frequency, WAL segment size, and bgwriter to maximize write throughput

Build a data-driven capacity planning process using historical metrics, growth trends, and resource modeling

Evaluate Google AlloyDB — columnar engine, read pool nodes, ML integration, and comparison with Aurora PostgreSQL

A deep dive into PostgreSQL JSONB vs JSON storage formats, GIN and B-tree indexing strategies, containment operators, and the performance characteristics that determine which to use for document storage workloads.