Context
A leading international provider of communication, public relations, and media intelligence solutions delivers a SaaS platform designed for marketing and communication professionals. Its offering combines media monitoring (press, TV, radio, podcasts, social media), campaign and media contact management (journalists, bloggers, influencers), and advanced analytics and reporting tools.
At the heart of this platform lies a critical customer portal, built on a microservices architecture and running on a Docker Swarm on-premise cluster (VMWare). The application comprised 30 microservices and relied on a MarkLogic NoSQL database of nearly 3 terabytes, essential to ensuring service continuity. Used daily by thousands of end customers, the portal required maximum availability, as any downtime could directly impact the user experience and client satisfaction.
To meet these challenges, the company decided to migrate its infrastructure to the cloud. The objectives were to:
- Gain scalability, resilience, and flexibility.
- Modernize the technology foundation.
- Ensure a seamless migration with minimal downtime.
Accomplishment
-
The migration was carried out progressively, in several key steps:
- Infrastructure Preparation
- Establishing a high-bandwidth, low-latency interconnect between the on-premise environment and Google Cloud Platform (GCP), ensuring smooth and secure data flows.
- Microservices Migration & Network Rationalization
- Gradual deployment of 30 microservices on Google Kubernetes Engine (GKE), while keeping the MarkLogic NoSQL database on-premise during the initial phase.
- Converting Docker Swarm configurations into Kubernetes manifests using Kustomize, improving standardization and maintainability.
- Migrating CI/CD pipelines from Jenkins to GitHub Actions, streamlining workflows and reinforcing automation.
- Replacing physical F5 load balancers with native GCP load balancers, reducing complexity and improving cloud integration.
- Database Migration
- Migrating the 3 TB MarkLogic database to GCP using real-time replication, enabling a smooth cutover with minimal service disruption.
- Infrastructure Preparation
Outcome
✅ Critical availability preserved: the customer portal remained accessible throughout the migration, with near-zero downtime.
✅ Modernized infrastructure: transition to Kubernetes on GCP, ensuring resilience and elasticity.
✅ Enhanced automation: CI/CD pipelines modernized with GitHub Actions, reducing deployment times and simplifying maintenance.
✅ Infrastructure rationalization: physical load balancers replaced by GCP native load balancers, improving efficiency and flexibility.
✅ Future-proof scalability: a flexible architecture ready to handle traffic peaks and support future growth.
By migrating its 30 microservices and nearly 3 TB of data to the cloud, the company strengthened its operational reliability, safeguarded the experience of end customers, and accelerated its capacity for innovation.
Written by Jean-Philippe ROBBE, jprobbe@penonpartners.com Practice Leader – IT Operations Excellence at Penon Partners
Thank you for your active reading – Discover our services and why Penon!
