Reduce, reuse, recycle: the three Rs of cutting down waste benefit us all, but is it possible to make them fun, rather than a chore? Sharing platform GEEV believes so, as do its 350,000 unique monthly users. Established in 2016, the French company aims to reduce waste by encouraging people to exchange their preloved items on its platform, adding a social element that makes recycling more fun.
« We currently have around 1.2 million customer accounts and more than 800,000 objects available on the site, » explains Kevin Tran, CTO at GEEV. « We’ve grown very quickly by taking an app-first approach to make it really easy for people to upload items and build playful elements into the process. »
For example, the interface incorporates game-like elements: to respond to an ad they’re interested in, users have to give away a banana from their bunch, that is, use one of their credits. These free credits are renewed every day, although users can access additional credits either by buying them or giving away items.
What attracts customers to GEEV is its ease of use. At the beginning, we chose to focus on building native mobile apps to deliver the best possible performance. Now, we need a backend that supports our growing traffic, both from our apps and the website.
— Kevin Tran, CTO, GEEV
In order to handle its growth, GEEV wanted to make its containerized infrastructure more robust and reduce time spent on maintenance. Following a series of outages, it decided to move to Google Cloud to achieve those goals.
« What attracts customers to GEEV is its ease of use, » explains Kevin. « At the beginning, we chose to focus on building native mobile apps to deliver the best possible performance. Now, we need a backend that supports our growing traffic, both from our apps and the website. »
USING KUBERNETES FOR EFFORTLESS SCALING
When an app is designed to make a specific process easier, such as ordering something or giving it away, users quickly become frustrated if they’re faced with long processing delays. « As the number of requests began to grow, our previous provider became less able to handle the volumes of traffic, » says Kevin. « We were suffering various outages ranging from a few minutes to several hours on a regular basis, and as we don’t have any full-time operations staff, it was up to me alone to resolve those issues. » Kevin began looking for a way to make the infrastructure more robust and ease his operations burden.
With support from implementation partner Op-Rate, GEEV migrated its infrastructure to Google Kubernetes Engine (GKE) on Google Cloud. Op-Rate also helped Kevin choose the best suitable Google Cloud services for GEEV’s platform and needs. « As Kubernetes originated with Google, GKE on Google Cloud doesn’t require any custom integrations, and this was a big factor in our decision, » says Kevin. « I also found the interface more intuitive, which is important when you’re trying to save time. »
Over the course of a day, our traffic might multiply by four or five times. Without the autoscale feature in Google Kubernetes Engine, it would be time consuming to handle and take up 25 percent of my time. I now spend only an hour or so per week on Kubernetes monitoring.
— Kevin Tran, CTO, GEEV
After a month spent perfecting the infrastructure, Op-Rate managed the staged migration over a two-week period, beginning with a testing environment and then sending a third of production requests to Google Cloud, to make sure the new infrastructure was stable. The change-over itself required zero downtime.
One of the key reasons for GEEV’s choice of Google Kubernetes Engine is its autoscale feature. « I simply set up the minimum number of instances and the CPU threshold, and it handles everything else for me, » says Kevin. « Over the course of a day, our traffic might multiply by four or five times. Without the autoscale feature on Google Kubernetes Engine, it would be time consuming to handle and take up 25 percent of my time. I now spend only an hour or so per week on Kubernetes monitoring. »
REDUCING OPERATIONAL WORKLOADS
With GKE in place, GEEV is expanding its use of Google Cloud tools to lighten its operational load. It uses Firebase for both its mobile applications and to host its web app, including Firebase Crashlytics for easy troubleshooting. To make data analytics more accessible, it is also developing a project with BigQuery.
« By synchronizing our hosted database with BigQuery, we will be able to serve data straight into Looker Studio, » says Kevin. « That means in-house teams will be able to access the numbers they need themselves, without having to ask us to generate reports. We can also use it to filter the analytics we deliver to our partners. »
Because we don't have to worry about the infrastructure anymore, we can really focus on the code we're producing. As a development team, we're able to concentrate on features, and this ultimately leads to a better experience for our customers
— Kevin Tran, CTO, GEEV
GETTING READY FOR TRANSATLANTIC GROWTH
Since migrating to Google Cloud, GEEV no longer experiences outages or downtimes. « It’s been six months since we migrated to Google Cloud, we haven’t experienced any downtimes from infrastructure failure, which is great. I can sleep better at night, » says Kevin.
Because the team is spending less time on dealing with system maintenance, they are able to focus on other areas. « Because we don’t have to worry about the infrastructure anymore, we can really focus on the code we’re producing, » says Kevin. « As a development team, we’re able to concentrate on features, and this ultimately leads to a better experience for our customers. »
To continue optimizing, as well as Compute Engine and App Engine, Kevin is focusing on implementing more Google Cloud tools that support automation. « We’re planning to use Cloud Storage together with Cloud Pub/Sub to automate data exports, » he says. « We’ll also use Cloud Functions to handle user image uploads on the app, which will make the user experience much faster and smoother. »
In addition to growing its market in Europe, GEEV’s next step is to expand to North America. « We want to take the GEEV revolution to the U.S., » says Kevin. « With Google Cloud, it’s going to be easy to set up a more dispersed infrastructure and use multi-region clusters to handle our U.S. and Canadian traffic. That’s really important, as we want to continue delivering the same excellent user experience as we grow. »
