The JusDB Database Engineering Blog

Redis Sentinel and Redis Cluster both provide HA, but teams regularly choose the wrong one. Sentinel is simpler and covers 95% of use cases; Cluster is required for horizontal scale. Here's when to choose each.

Use MySQL 8.0 generated columns to index computed values, JSON paths, and date parts. Covers virtual vs stored tradeoffs and functional index patterns.

Survey essential PostgreSQL extensions: pg_stat_statements for query analysis, pg_trgm for fuzzy search, pg_cron for scheduled jobs, pg_buffercache for cache inspection.

PostgreSQL's 5-10MB connection overhead means 500 app servers exhaust your database before a query runs. pgBouncer 2.0 multiplexes thousands of connections into dozens of server connections with improved auth_query and TLS.

Cut AWS database costs systematically: find over-provisioned RDS with Compute Optimizer, stop dev DBs after hours, buy Reserved Instances, and control storage growth.

Debezium 3.0 turns your database transaction log into a real-time event stream with full before/after row images. Learn to configure CDC for PostgreSQL WAL and MySQL binlog, avoid the offset storage trap, and monitor replication lag.

Guide MySQL query execution with optimizer hints. Covers INDEX, NO_INDEX, JOIN_ORDER, MAX_EXECUTION_TIME, and query block naming for subquery hint targeting.

Use PostgreSQL CTEs and recursive queries for category trees, graph traversal, and complex aggregations. Covers MATERIALIZED vs NOT MATERIALIZED and cycle prevention.

Master MySQL 8.0 window functions for analytical queries. Covers ROW_NUMBER, RANK, LAG/LEAD for period comparison, running totals with SUM OVER, and NTILE quartiles.

The InnoDB redo log default of 100MB is dangerously undersized for production. Learn to measure your write throughput, size innodb_redo_log_capacity correctly, and monitor checkpoint pressure before write stalls occur.

Tune PostgreSQL 18 WAL and checkpoints without weakening crash safety: understand commit modes, measure checkpointer deltas, and validate archive recovery.

Build a production database DR runbook covering RTO/RPO definitions, failover steps, PITR recovery commands, and a verification checklist. Test it quarterly.