20 Critical Database Checks Before Going to Production

Essential checklist covering memory sizing, connection poolers, backup automation, and OS limits prior to any production deployment.

Prerequisites & Safety Rules:

SSH root or DBA privileged accessRead/write access to DB config filesBackup destination write permissions

OS Limits (ulimits & vm.max_map_count)

Ensure file descriptors (nofile) is at least 65535 and swappiness is set to 1 or 10 to avoid swapping active database pages.

ulimit -n 65535
sysctl -w vm.max_map_count=262144
sysctl -w vm.swappiness=1

Connection Pooler Enforcement

Direct high-concurrency client connections must be routed through PgBouncer, ProxySQL or HikariCP with strict client connection limits.

Automated Backup & Restore Drill

Backups alone are insufficient; automated restore verification to a test instance must be validated before go-live.

Slow Query Logging Threshold

Enable query profiling and log queries exceeding 250ms to catch unindexed production queries early.

# PostgreSQL:
log_min_duration_statement = 250ms
# MySQL:
slow_query_log = 1
long_query_time = 0.5

Community Benchmark & Contribution

Share Your Restore Drill Results or Experience

Contribute real-world RTO/RPO telemetry, PostgreSQL version, backup methodology, and obstacles encountered from your infrastructure to help build industry benchmarks.