Case Study: Plaid achieves improved Amazon Redshift performance with Periscope Data

Plaid - Customer Case Study

Plaid, a banking-industry customer of Periscope Data since 2017, faced volatile and often slow query performance from their Amazon Redshift analytics cluster as Periscope Data was rolled out company-wide. With hundreds of dashboards and users of varying SQL skill, poorly configured Redshift tables (missing sort/distribution keys), unmanaged workload settings, and a flood of raw logging data, Plaid’s internal analytics were producing long-running, duplicative queries that hindered adoption.

Using Periscope Data’s usage meta-tables to pinpoint problem queries, Plaid rebuilt its analytics around rollup tables, materialized views and pre-segmented tables (materialized via Airflow), adjusted sort/distribution keys, and improved WLM/VACUUM practices. With Periscope Data enabling targeted migration and monitoring, Plaid removed nearly all “unbearably long” queries and achieved a dramatic performance gain—migrating 10x the number of queries while cutting query runtime to roughly 1/10th.

Plaid