GCP – repareo adopts cloud-based microservices architecture to scale its auto repair business

When it comes to auto repair, most of us are at the mercy of our mechanic. We take our car to the local garage, pay for the job, and hope for the best. And with little transparency into the work that has been done, or the costs involved, we can be left feeling unsure whether we got value for money.

At repareo, we are transforming the customer experience of vehicle repair, modernizing the market to make it e-commerce ready, while giving customers full transparency over the car-repair process. Now, they can describe their vehicle problem on our site and immediately receive a list of local garages, including customer reviews, cost breakdowns, and availability, allowing them to make an informed decision about where to take their car. And because our site has direct interfaces with the garages’ booking systems, customers are able to book their preferred garage and appointment directly on our site, saving time calling garages for availability. 

Outgrowing our monolithic architecture

This year, we will be launching our new infrastructure on Google Cloud. Previously, repareo was built on a monolithic system using a small, hosted server, which was both easy for our small development team to maintain and allowed us to grow the business in a cost-efficient way. However, as we added more features and services, our monolith grew, which had an impact on our development speed, creating a bottleneck for the rest of our application. 

Reliability became an issue too. repareo is integrated with car management fleets and leasing companies, and many drivers access our services through their car leasing app, making us highly dependent on third-party APIs to function effectively. As we grew, the increase in traffic resulted in these APIs becoming sluggish. During periods of peak traffic, such as the German tire-change season, we would see a 300% surge in traffic, placing a significant strain on our server, which was unable to scale effectively, causing our services to grind to a halt. 

Modernizing infrastructure to modernize the market

The turning point came last year, when we signed a major deal with a leading global e-commerce player to integrate with its vehicle parts marketplace, enabling customers to book an installation during the checkout process. Realizing that we would need our infrastructure to be able to handle an expected tenfold increase in demand, while conforming with our new service-level agreements (SLAs), we decided it was time to move to the cloud.

We knew that a migration would bring other benefits too, enabling us to build a microservices architecture to develop services in modules, as well as allowing us to place certain workloads close to a leading global eCommerce player in vehicle parts in California, for fast response times. As we looked at cloud providers, Google Cloud stood out for the range of technologies and services it offered, with BigQuery and Apigee particularly impressing us as uniquely advanced solutions in the market.

Beyond the technology, however, we were just as impressed with Google Cloud’s deep understanding of our industry and its business network within the automotive sector, as well as by the personal relationship we quickly built with the Google Cloud team. Migrating to a new provider is a once-in-a-lifetime decision, a marriage of sorts, and with Google Cloud the relationship immediately felt like it was built to last. 

A robust, reliable system for a smoother customer experience

We’re currently halfway through our migration, which has proved to be a steep learning curve for us, given the scale of the undertaking for our small team of developers. However, Apigee has helped to make the migration smooth by enabling us to easily set up a staging environment to test and adjust our system before going live, with no impact on our users. 

We expect to have completed the migration in less than six months in total. Once live, we will have a robust, scalable system, capable of meeting the needs of a significantly larger user base. Building and managing our APIs with Apigee means we will be able to use the caching system to cache the high number of API requests on the site, allowing us to offer high-performance buffering algorithms without having to drastically increase the scale of the underlying system. And because Apigee’s logging system is so well developed, we will easily be able to spot and remedy any integration issues, to ensure our APIs function effectively. As a result, our customers will enjoy a smooth, reliable booking system and real-time repair updates, while garages will benefit from far wider reach. 

We won’t need to worry about being able to handle fluctuations in demand either, as the autoscale feature of Google Kubernetes Engine (GKE) will automatically scale up our workloads to meet surges in traffic and scale down again during quiet periods to help ensure we aren’t using more compute than necessary. This cost-efficient provisioning means we will no longer have to worry about the ability of our system to handle periods of peak traffic, with our customers benefiting from a fast, reliable service. 

Taking our developers up a gear with easy-to-use, managed services

Our development efficiency has already significantly improved, thanks to the built-in features and managed services of Google Cloud. Apigee, for example, has a built-in key management system to enable APIs to communicate securely, which means our development team doesn’t need to spend time and money building our own system. Similarly, the fact that Cloud SQL is a managed database means we don’t need to spend time updating and maintaining it. GKE, meanwhile, improves our developer efficiency thanks to its easy integration with automation tools, increasing our deployment speed by at least 15%.

All of this means that our developers are free to focus on the core business logic and developing new features, such as new RPA-based technology to gather appointment availability from garage websites, which we were able to build and release inside a week, where it would have taken a month using the old infrastructure.

A firm foundation for sustained growth

With a number of other significant deals in the pipeline, repareo is now entering a period of sustained growth as we rapidly increase our customer base and prepare to enter new markets. That level of scaling simply wouldn’t be possible without Google Cloud, with its global network and regional data centers making it easy to move into any new region and enjoy rapid response times. While its scalable architecture means we can be confident that our infrastructure will always be able to scale with us.

As we continue our mission to make auto repair more transparent and convenient for more people, we are confident that with Google Cloud we have the right provider to help move our business into the fast lane.