Our client is an online grocery service that delivers to residences and offices throughout major metropolitan areas on the Eastern U.S. Seaboard. As its trusted database managed service provider since 2005, Pythian supports the client’s Oracle- and SQL-driven data estate from end to end and has helped the company optimize data operations several times. In 2015, for example, Pythian made application and database changes that improved customer log-in times by 120X. And alongside our business partner, Miro, Pythian in 2019 optimized the client’s Oracle licensing to the tune of $700,000 in savings in just a few weeks.
Then, in early March 2020, as the Covid-19 pandemic descended upon the NYC metro area and non-essential businesses were shuttered, this grocery service experienced an unprecedented surge in online orders. A system that previously processed around 12,000 orders per day was suddenly dealing with around 18,000 per day; in a matter of hours, the company’s SQL server environment was facing a massive 38,000 transactions per second. The company needed to quickly scale its systems to meet this sudden demand while maintaining its target customer metrics, ideally with no new CapEx.
Along with this explosion of new and returning customers in a very short timeframe, the amount of goods per order also increased by an average of 5X per order compared to historical averages – a potent double strain on the company’s systems that ratcheted up in a matter of hours and just kept on growing. Pythian DBAs jumped on the issue immediately, working around-the-clock for several days to stabilize the system and avoid an outright crash. The Pythian DBA team then spent the next six weeks reviewing, tuning, and optimizing the database in both its internals and operating system layers in an iterative fashion, while also changing some of its internal mechanisms to reduce server load and improve responsiveness.
Pythian DBA experts achieved an order of magnitude improvement in the performance of its Oracle and SQL environments, ensuring the client maintained a desired baseline page load time of three seconds with no new hardware investments (allowing the company to avoid more than $90,000 in potential CapEx, along with potential supply chain issues while they waited for new equipment to arrive). By mid-May, the client servers had been tuned and optimized to such a degree that loads effectively mirrored those of pre-coronavirus days – despite the continuing high demand for the client’s online grocery services.
PYTHIAN®, LOVE YOUR DATA®, and ADMINISCOPE® are trademarks and registered trademarks owned by Pythian in North America and certain other countries, and are valuable assets of our company. Other brands, product and company names on this website may be trademarks or registered trademarks of Pythian or of third parties. Use of trademarks without permission is strictly prohibited.