Imubit GCP Use Case

About Imubit

Imubit provides an AI process optimization platform for refiners and chemical operators, which executes plants’ money-making operational strategies continuously and consistently. The platform interconnects various processing units while aligning planning and economics, process engineering, process control, and operations for closed-loop optimization.

The Challenge

Imubit began its activity around 2016 and has been growing since then, they choose GCP for their application from the beginning. Trying to be cloud-agnostic they did not use Cloud native solutions and built all on IaaS. But, with this growth came a need to re-examine the environments,  tools and technology being used, and Imubit’s ability to efficiently control costs, maintain environments and create new Dev applications. Current application stack will be modernized to cloud-native technologies and tools like Docker, Kubernetes, Terraform, CloudSQL.

The Solution

Imubit choose GKE for running microservices in production and Dev (implemented with Namespaces), for monitoring and visibility was chosen Prometheus – Cloud-Native monitoring for Kubernetes & VMs with exporters for all kinds of backends used e.g., CoudSQL and  AlertManager for alerts routing and deduplication of  Prometheus alerts. In order to reduce maintenance and improve security CloudSQL was used as a MySQL on VM replacement. Kubeflow was used to implement machine learning workflows and improve ML CI/CD processes.

Cutting-edge technologies – consulting on and delivering a variety of workloads, taking into account Kubernetes and Istio to improve security and cost-effectiveness.
Security compliance – consulting on selection of third-party tools to support Imubit’s Security requirements, including endpoint protection, central log management, application audit logs, data protection algorithms, Kubernetes security, infrastructure perimeter security, and evidence provided as Infrastructure as Code (IaaC).

The results

Modern, secure, and elastic infrastructure on GCP with efficient cost and low operational cost. Visibility on traffic using GCP LB  and monitoring using Prometheus and Grafana dashboards.  IaaC and CI/CD pipelines used to onboard new workloads an easy thing.