TAPUZ - SQL Optimization at Scale
Israel's largest shipping platform - read the EXPLAIN plan, fixed the index, dropped p99 from 50s to 4s.
TAPUZ - SQL Optimization at Scale
The work
Resolved 90+ open production tickets for Israel's largest shipping e-commerce platform (NuxtJS frontend, NestJS backend).
Analyzed execution plans, identified full table scan across ~2M rows, validated cardinality estimates, introduced composite indexes aligned with JOIN/WHERE patterns, restructured subqueries, implemented keyset pagination, and added Redis caching with explicit invalidation rules.
Every change was benchmarked before and after. Indexes that didn't move p99 came back out before the PR landed.
Outcomes
- Query time improved from over 50 seconds to under 4 seconds (92% improvement)
- Timeout errors eliminated
- Peak traffic capacity increased by 5×
- Database CPU and I/O load significantly reduced
Have a project like this?
Start a project