From hand-built servers to ten deployments a day
Hotel Spider keeps hotels connected to more than 400 distribution partners. Here is how its team retired 2006-era manual infrastructure and got to ten deployments a day without adding anyone.
Industry
Hospitality tech
Headquarters
Switzerland
Scale
400+ distribution partners
Runs on Cycloid
Infra-as-code, pipelines, InfraView
The outcome, in short
5 to 10
deployments a day, fully automated
5 h/week
back, no more releases by hand
0
manual steps from merge to production
Who is Hotel Spider
Hotel Spider is a Swiss hospitality connectivity platform. It sits between hotels and more than 400 distribution partners, Booking.com and Expedia among them, keeping availability, rates and bookings in sync. When the platform is down, its customers lose reservations, so reliability is not a comfort metric. It is the product.
The challenge
The infrastructure carrying that responsibility dated back to 2006. Resources were created by hand, deployments ran manually and without tests, and nobody had a full view of what was running where. Each release was an after-hours exercise in caution: possible, but slow, and dependent on the person doing it.
Growth made the arithmetic worse. Staying lean and profitable meant scaling the platform without scaling the team that carried it, and the manual way had no room left.
“We had several challenges: our move to infra-as-code, automation of our deployments, management of multiple environments, and finding a holistic view of our infrastructure.”
Yannick Blondeau, CTO, Hotel Spider
Why Cycloid
The team rebuilt its delivery path on Cycloid rather than assembling one tool per problem. Infrastructure moved to code from day one, so every environment is explicit and reproducible instead of living in someone’s head. Deployments run through an automated, tested pipeline that behaves the same at 3pm and at 3am.
InfraView and the pipeline overview closed the visibility gap: one place to see what is running, where, and in what state, instead of answering those questions by digging.
The outcomes
The headline number is throughput: 5 to 10 deployments a day, where releases used to be occasional, manual events. That pace hands back around 5 hours a week that used to go into doing releases by hand.
The quieter outcome is nerve. With human error out of the deployment path, the team trusts its own releases enough to ship at the end of the week, not just at the start of it.
“I’d love to compare the number of deploys made on a Friday night before platform adoption, and after. I’d say it would be bigger by an order of magnitude.”
Yannick Blondeau, CTO, Hotel Spider
What's ahead
With delivery automated, the roadmap opens up: breaking the 2006 monolith into microservices and moving more of the platform to the cloud, on the same reproducible foundation.
See it yourself
The platform behind this story
A 30-minute walkthrough of Cycloid on your own use case, no slideware.
See what your team could ship with Cycloid
A 30-minute demo on your own stack and use case. Bring your hardest environment.