Swiss Insurance Agency Drives Efficiency and Stability with Scalable GitLab on EKS

Introduction

In the fast-paced world of insurance, where customer satisfaction hinges on swift response and reliability, a Swiss Insurance Agency faced a critical challenge: their existing GitLab infrastructure struggled to keep pace with the demands of their agile development workflows. Manual scaling of runners led to bottlenecks and delays, hampering collaboration and efficiency. To unlock continuous delivery and adapt to fluctuating workloads, the agency embarked on a journey to design a highly available and scalable GitLab implementation on Amazon Elastic Kubernetes Service (EKS).

Problem

The agency’s GitLab environment faced several limitations:

• Limited Scalability: Manual runner provisioning was time-consuming and inflexible, hindering the ability to handle fluctuations in development workload.
• Downtime Risks: The lack of redundancy and fault tolerance increased the risk of outages, disrupting the development process and impacting customer service.
• Inefficient Collaboration: Bottlenecks caused by slow build times and pipeline execution delays hampered team productivity and collaboration.

To address these challenges, the agency sought a solution that would:

• Automate runner scaling: Dynamically adjust the number of runners based on demand, ensuring efficient resource utilization and optimal performance.
• Enhance Availability: Design a resilient architecture with redundancy and failover mechanisms to minimize downtime and maintain service continuity.
• Boost Collaboration: Streamline workflows and reduce build times to empower developers and optimize team productivity.

Solution

The agency implemented the following key solutions:

• Automatic Runner Scaling:

1. 1. 1. Leveraged the Horizontal Pod Autoscaler (HPA) to automatically scale runners based on CPU utilization, ensuring optimal resource allocation.
      2. Defined custom metrics and thresholds to trigger scaling based on specific workload requirements.

• Resilient Architecture:

1. 1. 1. Deployed GitLab on EKS using a multi-AZ (Availability Zone) configuration for redundancy and fault tolerance.
      2. Implemented PostgreSQL HA cluster with replication for database high availability.
      3. Utilized Kubernetes readiness and liveness probes to ensure automated pod restarts in case of failures.

• Continuous Delivery Pipeline:

1. 1. Integrated GitLab CI/CD with automated testing and deployment processes to streamline workflows and reduce manual intervention.
   2. Optimized build scripts and container images to improve build times and pipeline execution efficiency.

Results

The implementation of this scalable GitLab infrastructure on EKS yielded significant benefits, including:

• Improved Scalability: Automatic runner scaling seamlessly adjusts to workload fluctuations, eliminating bottlenecks and delays.
• Enhanced Availability: High availability architecture minimizes downtime and ensures uninterrupted service even during potential disruptions.
• Increased Productivity: Streamlined workflows and faster build times empower developers and boost team collaboration.
• Faster Time to Market: Continuous delivery pipeline facilitates rapid deployment of new features and updates, improving responsiveness to customer needs.
• Reduced Costs: Efficient resource utilization minimizes infrastructure costs and optimizes cloud resource spending.

Conclusion

By embracing a scalable and highly available GitLab infrastructure on EKS, the Swiss Insurance Agency has transformed its development process. Automation, resilience, and efficient workflows empower their teams to deliver exceptional customer service and drive continuous innovation. This case study is a testament to the power of cloud-native technologies in unlocking agility and efficiency in today’s competitive insurance landscape.