🚚 White-Glove Heroku migration service. Learn more.

YC Series Part I: How Avenue scaled after the batch without hiring DevOps

How Avenue scaled after YC without hiring DevOps

YC Series Part I: How Avenue scaled after the batch without hiring DevOps

A Classic Dilemma: Speed or Scalability?

A year ago, Avenue exited the YC W21 batch faced with a hallmark problem for successful early-stage startups. Fresh off their seed round, Avenue was quickly picking up steam as more operations teams began to adopt their observability platform. Accommodating more users, however, meant taking a hit to product velocity to invest in DevOps and infrastructure.

Avenue's observability platform lets ops teams monitor and respond to business-related alerts

As a business intelligence tool, Avenue's platform also needed to connect to each client's production database, so technical complications (like requiring a static egress IP for users to whitelist) began to rapidly accumulate early on.

"After YC we were forced to start thinking about scaling things up, but we didn't want to have to choose between performance and features. We really wanted to just keep focusing on velocity and product."
Jeff Barg, CTO of Avenue

Unfortunately, these requirements were outside the scope of Avenue's existing hosting setup, which relied on Render and multiple AWS EC2 instances to run application workloads. In order to continue onboarding larger users, the Avenue team had a clear set of hosting requirements coming out of YC:

  • A Heroku-like experience for developers with minimal DevOps overhead
  • Infra that could scale to millions of requests per hour in a cost-effective manner
  • Networking customization for securely connecting to customer databases

As Jeff, Avenue's CTO, recalls: "We wanted something future-proof we wouldn't have to move off of. Many of our users need to whitelist our dashboard to connect to their internal services, so another migration down the line is something we wanted to avoid at all costs."

Moving to Porter: Velocity Without Compromise

After running into various networking and scalability limitations during YC, the Avenue team initially explored building their own deploy pipelines within AWS, testing an assortment of CI and IaaC tools including Waypoint and Terraform. Ultimately however, they found that their developer experience notably suffered, and AWS resources in particular were poorly documented as a whole.

"In the end, we decided not to burn several weeks of time on a homegrown setup. There’s a million ways to screw things up and only a couple ways to do it right. No matter what, it's not a situation you want to be in especially as an early-stage founder."
Jeff Barg, CTO of Avenue

Re-approaching the situation with an eye towards using an existing service instead of building something from scratch, Avenue discovered Porter and got everything up and running within two days.

Now on Porter, every developer at Avenue can easily manage and deploy services to the company's own AWS cloud, and key issues like static egress IPs for users to whitelist are handled out-of-the-box.

In addition to delivering value for core application and worker services, Porter has also enabled the Avenue team to seamlessly deploy and manage add-ons for services like Redis, Datadog, Metabase, and Tailscale.

Scalability for the Long Run

By delivering the convenience of a complete app platform in Avenue's own AWS environment, Porter has allowed the Avenue team to optimize developer experience on top of the full scalability of Kubernetes. Beyond the traditional benefits of a PaaS, Porter also brings tighter integration with native AWS services like RDS and ElastiCache - all securely connected via internal traffic to Avenue's Porter-managed workloads on EKS.

"Porter gave us Heroku-style automation on top of all the benefits of hosting in our own AWS, all without having to hire a single DevOps engineer. Kubernetes is purely an implementation detail on Porter, but it's also been great to tap into the existing ecosystem of vendors and add-ons."
Jeff Barg, CTO of Avenue

Thanks to this automation, the engineering team has been able to avoid dedicating technical bandwidth toward generic infrastructure work as they continue to scale. Instead, Avenue's team can keep focusing on high-leverage product improvements and new features to continue delighting their users.