Turo’s Car Sharing Service Iterates Quickly and Tests Often
New Relic sponsored this podcast.
Iterate quickly and let the product team test out ideas, says Adam Bovill, director of engineering at Turo, the car-sharing service that connects owners with travelers.
“If you can test ten, 15 or 100 ideas quickly and effectively you can find one, two, or ten that are good,” said Bovill, adding that as the Turo team is working through ideas, its data science team comes up with different models to understand risk. It uses the New Relic APM extensively, with New Relic doing 99 percent of its monitoring.
Lately, the company has started using Kubernetes and Spinnaker.
“Our goal right now is to accelerate development velocity. We’re not running into technical performance problems at the moment, and our goal is to figure out how to let developers develop as fast as they can develop. Kubernetes lets developers push their code out. We’re using Spinnaker as our deployment orchestrator but we have a model that developers push the deploy button, they write their code and when they’re done, it gets reviewed, they push the button that makes it go live. That’s a general principle across our entire team, and making sure everyone has the ability to understand that it’s gone out successfully, and if there are problems to understand where problems happen,” said Bovill.
In this Edition:
5:47: What is the value that Kubernetes provides underneath everything?
7:54: Tell me about the teams, and how you’re thinking about team development in terms of how you’re building out your microservices story.
14:14: Load balancing and resource management in Turo’s monitoring environment
19:11: How many New Relic APM clusters are you using right now?
23:32: What are some of the patterns you can see building upon Kubernetes and New Relic APM?
26:07: Are you building new applications on top of Kubernetes, or are you mostly using it for other purposes?
Feature image via Pixabay.