The Challenges

Stumbling Blocks for the Customer

• 

  Monolithic System

  The customer was operating on a monolithic system, challenging to identify which subsystem or upstream service was failing.

• 

  Struggles with Testing

  The developers often faced difficulties in testing multiple services simultaneously. In addition, with many moving parts, it was troublesome to trace any errors.

• 

  Scaling and Cost Issues

  They encountered scaling challenges and a sudden rise in infrastructure spending; they needed specialized consulting to optimize their costs and performance.

The Cloud

Why Amazon Web Services?

The customer was already operating on Amazon Web Services. However, CloudCover introduced/proposed certain AWS services bundled with third-party tools & Microservices model to increase scalability and improve performance.

• 

  Powerful Microservices

  Amazon’s ECS and Spotinst (a third-party tool) helped break the monolithic architecture into microservices in a highly cost-effective way.

• 

  Dynamic Provisioning with Infrastructure as code

  AWS CloudFormation massively helped in creating reusable IaC. Deployments could be automated seamlessly and quickly across environments, thereby eliminating the need for manual intervention.

• 

  24*7 Monitoring & Automation

  CloudWatch helped in the monitoring of services and creating automation for self-healing systems.

The Strategy

Our Approach

• 

  Microservices Model Adoption

  The objective was to decouple their services and improve performance and monitoring. CloudCover also introduced a scalable solution that helped perform custom CI on the code base and allowed continuous deployments and delivery to various environments and infrastructure.

• 

  Powering GitLab

  By making the best use of the customer’s Gitlab that was already available for managing the code base, CloudCover utilized GitLab-runner to perform our CI/CD leading to more flexibility to implement the CI/CD.

• 

  Datadog for Observability

  CloudCover used CloudWatch Logs, Metrics, and Custom metrics to monitor the infrastructure and services natively available on AWS. With the help of Datadog, it was easy to monitor the application performance and collect traces to aid the developers in gaining more visibility on how the applications were performing.

• 

  Custom Solution for On-Demand Infrastructure

  The strategy was to integrate the existing infrastructure with scalable, low-cost, and HA solutions with an ever-increasing customer base. CloudCover provided a custom solution to the customer’s developers to create on-demand infrastructure with spot pricing. In addition, it allowed them to test multiple services/codebases simultaneously and delete the infrastructure built once testing was complete.

Technical Solution

Alleviating the Pain Points

• 

  Monolithic to Microservice

  • CloudCover used Gitlab’s native feature hosted GitLab-runner to perform CI/CD.
  • In addition, with AWS Cloudformation, CloudCover was able to implement Infrastructure As Code (IaC) and parameterize deployments on AWS Elastic Container Service (ECS) to reduce human intervention and errors, which helped break the monolith design and gave the flexibility to add more services and APIs to the infrastructure.

• 

  Monitoring and Better Analysis

  • Datadog agent was installed on each ECS instance to monitor the services and the APIs hosted on them. Thus, one could quickly gain insights into the performance of these services and the APIs were performing.
  • In addition, AWS CloudWatch was used for logging ECS Services to a central location for better filtering, analysis, and debugging.
  • Amazon’s APM heavily contributed to honing the applications in a better way.

• 

  Scalability with Spotinst

  • Moving the majority of the infrastructure to Spotinst and choosing multiple instance types made it possible to achieve high scalability, HA, and low cost.
  • In addition, integrating ECS Clusters with Spotinst to run the microservices and APIs reduced the cost massively.

Business Value Delivered

What We Accomplished

• 

  Microservices and 360° Observability

  CloudCover successfully moved the customer from a Monolithic architecture to Microservices; utilized ECS to its fullest potential while running the clusters on the spot to manage the cost. Today, the customer is running more than 15 services on ECS with minimal impact on their expenses. In addition, Amazon CloudWatch, alongside Datadog, has helped effectively monitor the infrastructure, applications, and APIs. As a result, the customer can now efficiently manage all the idle and unused resources by paying for their use.

• 

  Reduced Infrastructure Spend & Custom Script for Testing

  One of the critical challenges faced by the development team was regarding testing. With the Custom script, the customer’s developer team was able to run their tests in parallel without any overlaps with fellow developers. Also, running them on the spot decreased infrastructure spending. CloudCover and Spotinst drastically reduced the costs for the customer by up to 70% on EC2.

• 

  Scalable Infrastructure & Speedy Deployments

  CloudCover created a more scalable infrastructure to deploy any service to production within 30 mins with minimal human intervention. In addition, by introducing custom CI/CD and stages, the customer could monitor the code going to production, ensuring only validated code gets pushed forward.