About Barstool Sports
Barstool Sports (now a part of PENN Entertainment) produces content for distribution on multiple platforms and through different business models including web, apps, podcasts, radio, video, social, events, pay per view, subscription and television to an enthusiastic audience. After starting out in print in 2003, Barstool Sports followed the growth of online publishing and soon switched to video production.
The Challenge
Barstool Sports is one of the world’s leading digital media companies which manages a multitude of ever growing properties. Being able to efficiently serve content to users is paramount and Barstool leverages a number of service providers in conjunction with their AWS infrastructure to do this in a cost effective manner. Each property has a specific set of AWS and third party primitives that work in tandem together.
Ensuring these various services work together in a cost-efficient manner takes significant engineering time and analysis and the consequences of getting something wrong directly result in massive cost overruns and margin erosion.
The Solution
A single pane of glass view of cloud spend.
Barstool Sports uses Vantage to manage costs across AWS, GCP, Fastly, Datadog, Snowflake, and MongoDB Atlas. Vantage provides a single pane of glass to view all of the costs associated with each provider down to the individual resource level. Barstool Sports uses Cost Anomaly Detection to monitor and respond to spikes in CDN, monitoring, and data costs which come with their rapid growth and Barstool’s consistent deployment of new services and properties.
Fastly integration
As a media company, Barstool Sports pays careful attention to their spend on Fastly. They utilize almost every major service that Fastly provides, including CDN, Compute@Edge, Enterprise support, Image Optimizer, and TLS. Vantage provides resource level visibility into each of these Fastly services and enables Barstool Sports to view their spend alongside other services in Overview.
“Vantage has already helped us cut our total AWS bill by over 50%. When the Vantage team approached us about supporting additional providers beyond AWS, we were eager to get Fastly costs included. Previously we received the generic bill from Fastly on a monthly basis but were unable to easily see day by day costs on a per service basis. Vantage has saved us countless engineering hours by pulling relevant usage data from Fastly and cross-referencing our rate information from billing APIs to make available in the Vantage console. The best part is we can see Fastly and AWS costs in the same reports without having to log into multiple systems.”, said Andrew Barba, CTO at Barstool Sports.
Filtering by resource
Filtering costs by S3 bucket in the Vantage console.
Using Cost Reports Barstool Sports is able to filter their costs across providers by individual resources. This allows them to create reports which include multiple ECS clusters and one Fastly CDN to properly attribute costs to each project they are working on.
Data transfer charges
This resource-level support allowed Barstool Sports to reduce their data transfer charges significantly. For example, they made an optimization in a MongoDB Atlas cluster which was transferring data across AWS regions unnecessarily.
Datadog, Snowflake, and MongoDB Atlas
As Barstool Sports integrated Datadog, Snowflake, and MongoDB Atlas, the main concern was detecting anomalous costs on those services related to new deployments or spikes in traffic. In one case, there was a runaway Datadog charge associated with a new service on AWS Lambda that Barstool engineering was able to detect and remediate using Vantage. In other cases, the team has been able to spot Snowflake queries which were great targets for optimization because they accounted for a large chunk of costs.
Reducing Costs and Gaining Visibility
Barstool Sports now utilizes Vantage to solve all of its cost transparency challenges. With Vantage, Barstool Sports can easily see all of the costs associated with each resource broken down by cost category daily. The team can also associate AWS costs with accompanying third party costs, such as Fastly, Datadog, Snowflake, and MongoDB Atlas. Lastly, anomaly detection in Vantage helps Barstool Sports launch new services and products without fear of unexpected charges from the interactions of its multiple providers.
