Learn how the Czech branch of Delivery Hero moved from on-premise solution to fully managed Cloud. Revolgy provided a full Lift and Shift Google Cloud Platform migration at Dáme Jídlo in order to stabilise the infrastructure of a food delivery app, as well as accelerate deployment thanks to CI/CD pipeline optimisation.
Since 2012, Dáme jídlo has been helping people in the Czech Republic get the food they love, wherever they are, whenever they want it with just the push of a button. With more than 2,000 restaurants on its platform, Dáme jídlo was the first food delivery platform to provide its services to the whole country. The company puts particular emphasis on quick delivery and care for its users, but eventually its scale of operations began to strain the limits of its infrastructure, risking its reputation for premium customer service. When the time came to look for a new solution, Dáme jídlo looked to Google Cloud Platform (GCP).
“Our servers were unavailable for long periods of time, so we were losing orders and losing money,” says Petra Tříšková, Head of Development at Dáme jídlo. “We switched to GCP because we wanted to move to a robust, high performance solution.”
Dáme jídlo acts as an online marketplace, allowing customers to find restaurants and order whatever food they crave. In order to satisfy its customers as quickly as possible, it is vital for the company to have a smooth-running service with minimal disruptions. By late 2016, Dáme jídlo’s infrastructure was suffering regular outages, sometimes at peak hours. At the time, Dáme jídlo hosted its servers in a data center, which meant being removed from the servers themselves. “Whenever we had an outage, we had to raise a ticket with the center’s DevOps team to deal with it,” says Petra. “That took a lot of time and, ultimately, cost us money especially when the outages happened during lunch or dinner hours.”
Google Cloud Load Balancing enabled the website and mobile app to perform smoothly regardless of the load on the servers. While Google Cloud Storage established a secure way to store important static assets such as pictures or company documents. With Google Identity and Management, Dáme jídlo could extend and restrict permissions with ease, improving data security. Revolgy worked hand in hand with Dáme jídlo to help the company redesign its application architecture into a reliable distributed system leveraging cloud-native concepts and using containers on Google Kubernetes Engine.
This allowed for automatic up-scaling or down-scaling in response to traffic changes to maintain an optimal latency and billing cost.
Google Stackdriver Monitoring let the company gain an insight into what exactly was happening on their servers and enabled them to anticipate any problems that may occur. Revolgy then configured this setup with Slack and SMS for 24/7 alerting, that brought a new transparency to the infrastructure.
“The administration and billing side of things has really improved for me,” says Petra. “With GCP I can finally see how the servers are performing, how we’re using them and exactly how much each service costs.”
In addition to the added transparency, Revolgy also built Cloud Ops, a managed service that took care of all of Dáme jídlo’s infrastructure care from monitoring and alerting, to support, to Site Reliability Engineering. Revolgy and Dáme Jídlo simplified the developers’ workflow with a new Continuous Integration/Continuous Delivery (CI/CD) pipeline. This consisted of a fully automated software integration and delivery process that builds, tests and deploys their code every time there is a change pushed to their version control system. “As the infrastructure is declaratively defined (as code) in source control, rolling out a new version of the microservice causes no downtime for the app,” says Marek Klos, Project Manager at Revolgy. In addition, Google Cloud Deployment Manager provided comprehensive release management over the entire GCP infrastructure.
“Thanks to Google Cloud technologies, we were able to help Dáme Jídlo outsource their infrastructure operations completely,” says Marek, Project Manager from Revolgy. “Now they can focus on just building their product.”
With GCP and Revolgy, Dáme jídlo has built an infrastructure that helps it deliver the best possible service to its customers at all times. Unplanned outages have been reduced by more than 95%, according to Petra down from an average of more than one hour per month to only a couple of minutes. Google’s easy to use monitoring services combined with Revolgy’s Cloud Ops service gave Dáme jídlo an unprecedented level of transparency and management of its infrastructure, allowing the company to concentrate on its core business goals.
In addition to addressing the stability issues of the past, Dáme jídlo’s new Kubernetes-based architecture will provide an easy scalability as the company grows. “Migrating to GCP has cut our infrastructure costs as we no longer have to pay a data center for DevOps or extra monitoring,” says Petra. The company is currently working closely with Revolgy to identify more opportunities for efficiency with Google’s suite of cloud tools and create even more savings. For Dáme jídlo, the journey with Google is just beginning.
“Initially, I just wanted to make the servers more available and more reliable,” says Petra. “Now we are already in a phase where we can downscale and upscale the infrastructure based on our traffic. We’re so much further along than I expected!”