Room to grow.

01 / Context

The problem

Link, Kraken’s electricity settlement platform, works with large datasets. A view spanning approximately 26 million records was timing out, while redundant database indexes were consuming production storage.

02 / Responsibility

My contribution

I replaced expensive full-table counts with lazy pagination, added indexes for common filters, and removed redundant indexes across two production estates.

03 / Approach

The decisions

For the large data view, lazy pagination replaced full-table counts, and targeted PostgreSQL indexes accelerated common filters. Storage work addressed a different side of the same problem: identifying indexes that no longer justified their space. Adding an index can help a query; retaining every index indefinitely has a cost.

1. 01

   Find the bottleneck

   A data view spanning approximately 26 million records was timing out.

2. 02

   Change the work

   Replace full-table counts with lazy pagination; index common filters.

3. 03

   Recover space

   Remove 11 redundant indexes across two estates.

04 / Result

The outcome

The data-view timeouts were eliminated. Separately, safely removing 11 redundant indexes reclaimed approximately 1.43 TB across two production PostgreSQL estates.