How Arena Understood the Unit Economics of their Media Streaming Business is a cloud-based marketing platform for live experiences. Their mission is to empower media companies to take back ownership of their audience from social networks and grow their community on their website.

Arena enables publishers, media and broadcasting companies to cover events such as news, sports, finance and tech. Through a powerful AI technology, is capable to aggregate, curate and distribute content to any device, anytime, anywhere.

Why Arena Chose Serverless

Arena’s product is a real-time content platform. They cover live blogging for sports events. As part of it, there are many spikes in their traffic – during the Superbowl, for example. Serverless is great for this use case because it scales. There isn’t a lot of downtimes.

The Arena team has been using serverless technologies since 2016. They use many AWS services and APIs: AWS Lambda, CloudWatch, DynamoDB, Elasticsearch, Redis, and Elasticcache, in addition to external APIs such as Google, Twitter, Stripe, and more.

The Problem

The number one issue Arena had was blindness to what was happening inside their application. Since they started the serverless project, they had difficulties getting alerts about problems such as timeouts, errors, exceptions, etc. Trying to optimize Lambda functions and cost of their stack was also very challenging.

They used AWS CloudWatch and tried several other vendors. While being helpful, none of them helped them solve complex issues quickly.

Arena and Epsagon

Arena was looking for visibility into their distributed serverless application. The first benefit they saw in Epsagon was the tracing and easy-to-read diagrams.


“The onboarding was really fast”, Says Arena’s CTO, Rodrigo Reis. “By simply deploying Epsagon’s CloudFormation template we immediately saw functions timing out we didn’t know about, or other failing functions we weren’t aware of.”

“The tracing library was installed very quickly”, Says Rodrigo. “We saw graphs and diagrams right away. The team doesn’t want to spend time drawing these diagrams manually, it also takes a lot of time to keep them updated. Epsagon’s automated distribution helped a lot.

APIs performance

Besides helping them keep track and troubleshoot their own application, Epsagon also provides visibility into the behavior of vendor APIs such as Stripe, Google, and Twitter. It encourages accountability of vendors over the performance of their APIs and how it’s affecting their customers’ business.

Epsagon Compared to Other Solutions

Arena uses APIs such as Google and Algolia

Monitoring and logging in the same platform

There are some solutions out there that report the executions and exceptions of Lambda functions. However, every time there was an issue, CTO Rodrigo Reis had to go to the AWS console and check CloudWatch, which was very time-consuming. With Epsagon, Arena can get errors and logs in the same platform, which is “one step less” and saves a lot of time.

End-to-end visibility including 3rd party APIs

Before Epsagon, Arena didn’t have visibility of a complete transaction. For example API Gateway → Lambda → 3rd party APIs. Epsagon captures these distributed transactions automatically. The complete view of a business transaction helps to solve problems significantly faster.

Automated alerts

Previously, using other solutions, Rodrigo’s team had to manually create alerts based on thresholds, but they didn’t know how to set them. There was a lot of trial and error involved, consuming a lot of time. Epsagon alerts them about issues they didn’t anticipate automatically, without any manual management. Rodrigo is excited about the upcoming anomaly detection which is coming soon in Epsagon.

Benefits to Arena’s Business

Increased focus

Epsagon’s automated monitoring provides Arena’s team with increased focus and developer velocity. As a CTO, it helps Rodrigo manage his team and focus more on the business rather than spending time troubleshooting.

Better customer support

Using Trace Search, Arena is able to locate transactions based on specific customer ID. It is very helpful to solve support tickets quickly.

Major reduction in troubleshooting time

Compared to working with solely the AWS console, Rodrigo is seeing a total of 60%-80% reduction in troubleshooting time, helping the team focus on their business more than ever before.

Cost information and unit economics

Arena’s team has many Lambda functions and each of them is associated with a product. “Before Epsagon, I would lick my finger and point to the sky to estimate my serverless costs”, says Rodrigo. “It was mostly trial and error: playing around with the maximum duration, the code size, and the Lambda’s configuration. AWS only provides the overall cost.”

Using Epsagon, they now see the cost of each product. They constantly know the unit economics for each product and can prioritize the work on each product accordingly. This unit economics tells how much each feature costs, and see if they can optimize it or adjust its price. This leads to a simple ROI formula of their product.

The unit economics and the increased confidence helps the developer’s velocity and saves time for the team.

Decisions Made Following Epsagon

Design pattern change

Now that Arena has the performance of 3rd party APIs, they are changing the design pattern to concentrate all 3rd party requests in a separate function. This helps to throttle or queue them if needed, so it doesn’t overwhelm the 3rd party service.

Actions to save costs

Epsagon helped the Arena team reduce their AWS Lambda cost in several ways:

  • Based on the cost information, they split high-cost functions into decoupled smaller functions that cost less.
  • They fixed timeouts which cause functions to cost more than they should.
  • They optimized the memory of functions to reduce their costs.