Cloud Costs Optimization

What and why?

Udaan has grown from being a b2b e-commerce marketplace to being a logistics, warehousing, credit, data analytics company, and more. As a result, it houses more than 50+ databases catering to different domains.

Over the last 2 years, I’ve been involved in improving the state of databases at Udaan. We managed to cut down around 70% of the cost while catering to the increase in depth and breadth of the business. Doing that, I’ve had so many lessons and built diagnostics scripts & tools to run lean eventually.

Rough Agenda

1. Cloud offering specific optimizations - choosing the right offering
   Agenda: To help choose between serverless/provisioned, elastic pool, business-critical vs general purpose vs hyper-scale. Local disk vs network disk.
2. Role of database maintenance
   Agenda: Rebuilding, reorganizing, updating index statistics, redundant index analysis, and removal.
3. Understanding indexes and choosing the right one
   Agenda: IO vs storage tradeoff, clustered vs unclustered, order and specificity of columns, key lookups and included columns, conditional indexes, overlapping and duplicate indexes,
4. Leveraging Snorql and tools
   Agenda: How we took a tool-first approach to solve the above problems.
   We built a diagnostics and optimization tool to save our cost and time and decided to open-source this. And how to leverage this tool to do optimizations without being a hardcore DBA.