Case Study: Plaid achieves faster Redshift query performance and scalable analytics with Sisense for Cloud Data Teams

Plaid Uses Data to Manage its Amazon Redshift Performance

Plaid’s Data Science & Infrastructure team was charged with supporting a rapid rollout of Sisense for Cloud Data Teams on top of an AWS Redshift warehouse. As usage grew from power users to many casual analysts, query performance became volatile: tables lacked appropriate sort/distribution keys, workload management wasn’t tuned, and the cluster had become a dumping ground for raw logs. The result was inconsistent p90 runtimes and hundreds of slow dashboards driven by a small number of very large tables.

The team used Sisense meta-tables and Redshift diagnostics to pinpoint hotspots, then built pre-aggregated rollup tables, materialized views and pre-filtered derivatives, scheduled and swapped them via Airflow, and tightened maintenance and WLM settings. They rolled changes out dashboard- and user-by-user using tracking dashboards and cross-team collaboration; the outcome was a stabilized cluster with far fewer long-running queries, much faster dashboard performance, and broader internal adoption.

Plaid

Austin Gibbons

Software Engineer