What a Fractional DBA Actually Does for Your Team
Beyond the job title: the concrete weekly work of backup verification, upgrade planning, and incident response — without a full-time hire.
Bloat, stale statistics, and mystery slowdowns usually trace back to autovacuum settings written for someone else's workload.
PostgreSQL performance problems are frequently maintenance problems in disguise. The default autovacuum configuration is a compromise written for a workload you do not have. On a write-heavy table it falls behind; bloat grows; every index scan pays a tax; the planner reads stale statistics and picks worse plans.
A hot orders table and a cold reference table do not want the same vacuum cadence. Set scale factors and thresholds per table for the heavy writers; leave defaults for the rest. This one change resolves the majority of "PostgreSQL gets slow over time" reports.
Bloat estimates vary wildly between tools. Establish a repeatable measurement — dead tuples ratio, table size vs. live rows — and trend it. If autovacuum tuning is your fix, the curve should flatten within days; if it does not, the problem is elsewhere.
Beyond the job title: the concrete weekly work of backup verification, upgrade planning, and incident response — without a full-time hire.
Instance right-sizing, index pruning, and data lifecycle design — a savings playbook measured against latency, not hope.
Keep your flexible JSON schema and get index-speed lookups: generated columns turn hot JSON paths into first-class index targets.
Work with Burj
Tell us what is slow — CPU, queries, migrations, or cost — and we will scope the smallest safe fix.