/home

$ ls ./blog

Notes

Projects I’m building, ideas I’m testing, and thoughts from along the way.

Preventing Serial Processing on the Import Pipeline

The HubSpot import system is responsible for ingesting hundreds of millions of spreadsheet rows per day. We translate and write this data into the HubSpot CRM. In this post, we will look at how we solved an edge case in our Kafka consumer that was leading to poor import performance for all HubSpot import users.