Alerty Blog

Building A Fully Scalable AI-Powered Web App With A Sequoia-Backed CTO

Written by Jack Dwyer | Aug 23, 2024 3:30:00 PM

When building any application, especially an AI-powered one, having a robust and flexible deployment platform is crucial. We sat down with Zihao Zhang, CTO behind Sequoia-backed Alerty, to discuss how Alerty's engineering team is using Zeet to scale their operations and deliver a cutting-edge product to their customers.

Alerty's architecture is built on a foundation of microservices, designed to handle massive amounts of data ingestion, storage, and querying. The team knew from the outset that they needed a deployment solution that could support their event-driven, microservice-based approach while allowing for easy management of multiple services.

 

"We wanted to have a staging environment, we wanted to have branch deployments for the frontend to preview, we wanted to have microservice architecture, and we also wanted to have a data intensive back-end," explained Zihao.

One of the key requirements for Alerty was the ability to run staging and production environments side by side. This approach allows the team to thoroughly test changes before they reach production, reducing the risk of unexpected issues and improving overall reliability.

Integrating AI into the application presented unique challenges, particularly when it came to supporting streaming requests. The team developed a custom Go backend to handle streaming HTTP responses, ensuring that users can see results in real-time as they interact with the AI.

 

Zeet plays an important role in Alerty's development and deployment workflow. For the frontend, Zeet manages the CI/CD process, automatically deploying changes and providing preview URLs for pull requests and new branches.

"Zeet manages the CI/CD process and manages the preview environment, staging environment, and production environment," the team noted. "Our front end team is able to create that experience just from a few button clicks on the Zeet platform."

On the infrastructure side, Alerty utilizes several AWS managed services, including Kafka and Prometheus. Zeet's integration with these AWS services has been a boon for the team.

 

As Alerty continues to scale, the team is looking to leverage Zeet's multi-cluster deployment capabilities. While their current multi-regional deployments are orchestrated through Terraform, the team sees potential in using Zeet to manage multi-region deployments for their core application, including backend containers and workers.

One of the standout features of Zeet for the Alerty team is the visibility it provides into deployed resources. Unlike other deployment tools used in the past, Zeet offers an operational dashboard and control plane that allows developers to monitor the health and status of their deployments easily.

"Compared to other Platform-as-a-Service tools, the level of integrations it has with cloud providers is really useful to us," the team emphasized. "With Zeet, I have this native integration that lets me not only access these resources but also access them in a secure way through IAM access policies, and that's been a real game-changer for us."

 

 

In conclusion, Zeet has proven to be invaluable for Alerty as they build and scale their AI-driven monitoring and alerting platform. By providing easy integrations with cloud services, simple management of microservices, and robust staging and preview environments, Zeet enables the Alerty team to focus on developing innovative AI solutions for their customers.

As AI continues to reshape the tech landscape, platforms like Zeet are playing an important role in letting developers build, deploy, and scale complex applications with ease. For teams like Alerty, this means faster development cycles, more reliable deployments, and ultimately, a better product for their users.

If you want to catch the full interview, check it out here: