Porting from in-house datacenter to GCP of several workloads with use of managed services (Cloud SQL, Filestore, MemoryStore) as well as integration with data management tools (DataPrep, DataFlow, BigQuery), Kubernetes and containers.

The Company

iLMeteo is a technological company specialized in providing forecasting and weather news services for all Italian and worldwide cities.

The team, composed from about 20 people among executives, meteorologists, physicists, engineers and communication experts works 24x7x365 to guarantee updated and reliable forecasts.

The Challenge

The high workload during specific events (summer holidays, extraordinary weather events, etc.) caused periods of downtime for the website and the APIs. The previous infrastructure, on premise, outdated and based on physical servers, did not adapt to the current needs of scalability and resilience.

The Solution

The migration to the cloud (performed through the replatforming approach) allowed to achieve the scalability needed during the events. Nevertheless all the environments (Production, Development, Stage and Integration) were properly segmented and insulated.

With the migration, all the partner’s websites were also ported into containers on GKE.

The Project

After the initial lift and shift performed from the previous infrastructure to GCP, we orchestrated the transition and optimization from the legacy monolithic infrastructure to a microservices one in order to improve the availability, reliability and ensuring also a better scalability and cost optimization following the business growth.


The new infrastructure improved stability and reduced page load times, with a significant improvement in reliability and organic traffic on the site. 

Beyond this, the cloud infrastructure, being more stable, allowed a more efficient scaling during traffic peaks and a better management of any attacks.

The overall TCO was reduced compared to the previous infrastructure thanks to the consolidation of ht services and the optimized use of the GCP resources.

Compute engine enables us to autoscale instance groups during periord of low traffic, which means we now use a quarter of the numbr of instances compared to when we have higher traffic. That’s been a huge advantage for us
Emanuele Colli, CEO, iLMeteo

Recommended Posts