Achieved Milestones
- Migrated to Kubernetes in ~4 weeks
- Archived ~60% automation
- Increased collaboration between Developers and DevOps
- Time-to-market reduced by ~30%
- Developer Productivity increased by 40%
- Streamlined CI/CD for Kubernetes
- 4x Faster Onboarding of new applications Kubernetes
About The Company
Indiamart is one of the global leaders in the B2B online marketplace, connecting buyers and sellers. Currently, it is home to ~200 million registered buyers and ~8 million sellers listed on its marketplace. The company serves as a vital hub for business needs, focusing on creating value through powerful network effects, it simplifies business operations by providing access to a broader market, cutting-edge technology, and financial services. Through its innovative tools and services, the company simplifies the process for suppliers to establish and grow their online presence, driving success and growth for businesses of all sizes.
Background
As a global leader in the B2B online marketplace, Indiamart is heavily focused on cutting-edge technologies to support the millions of buyers meet the sellers with a wide range of tech stacks being used such as Apache, PHP, Java, Elastic Search, Kibana, Fluentd, etc. The company has been using a legacy tech stack, with a monolithic architecture, in-phase of transitioning into micro-services and deploying their application on Virtual Machines. With the increased growth and demand for scaling, Kubernetes is being introduced and planned to modernize their applications and deploy on Kubernetes.
Challenges Faced
The global B2B online marketplace is home to millions of users daily visiting the marketplace, thus having a huge volume of load that needs to be handled. Because of the nature of the complexities of monolithic architecture, it was difficult to manage the application deployments and scalability was a big concern. All the deployment patterns were manually executed which again caused delays in releases and if something went wrong, it was even difficult to troubleshoot, leading to downtime of the platform. Modernization of the application was required. Additionally, the time-to-market was very high due to the dependencies of the developer on the operations team for any kind of deployment/ testing. To solve these challenges, the team decided to move to the managed Kubernetes service provided by Google, i.e. Google Kubernetes Engine (GKE).
The team has explored a lot of different tools to streamline the CI/CD deployment process on Kubernetes, such as Jenkins, ArgoCD, and Kube-sphere, yet the problem remained the same. Here are some of the major challenges the team faced:
Dependencies on DevOps: Developers were still dependent on DevOps teams to deploy their applications on respective environments. Some level of automation was achieved with Jenkins and ArgoCD yet, the developers were not able to own the applications, especially for debugging and troubleshooting applications if anything went south.
Steep Learning Curve: The team decided to move to Kubernetes, but due to the nature of its complexities, it was difficult to onboard new developers on the K8s environment. The explored tools didn’t provide the kind of abstraction required for developers to work efficiently and since the team was moving onto Kubernetes, it was getting much longer time to onboard micro-services on Kubernetes as well.
Tool Chain Complexities: Streamlining the CI/CD workflows for Kubernetes was again a major setback. Using different tools like Jenkins, ArgoCD (one instance in one cluster), monitoring tools, etc are again increasing the complexity of the entire pipelines which was again another overhead for the team to manage these tools as well.
Access Management: It was important for DevOps teams to build fine-grained access control for developers so that they can access the applications deployed on Kubernetes in case any issues occur. But it was getting very difficult with the tools explored including Kube-sphere, and ArgoCD. Fine-grained access control was not available as per what they needed.
These were the major challenges that the team faced when they started moving to GKE, making it difficult to make their way into the Kubernetes ecosystem.
Solution: Devtron’s Software Delivery Platform for Kubernetes
In the Software Delivery Lifecycle, fragmented tooling can adversely affect the overall productivity and efficiency of the teams. The primary objective of the team was to streamline their CI/CD pipelines for Kubernetes with a platform that would boost developer productivity and make it easier for them to quickly onboard new applications on Kubernetes. During their explorations, Devtron came out to be the most ideal platform which solves all their use cases and provides a single pane of glass for all things Kubernetes.
Streamlined Kubernetes CI/CD Pipelines: Devtron comes with Kubernetes-native CI/CD pipelines that helped them to streamline the entire CI/CD process, and all under the same roof. Apart from the builds, Devtron does have DevSecOps capabilities which helped them improve the security posture, natively available within the pipelines. Additionally, gitops-based deployments with Devtron is a breeze. They didn’t have to worry about writing and managing the manifests files, and application objects which ArgoCD will fetch and deploy from git. Devtron automates the entire process.
Empowering Developers: The abstraction layer that Devtron provides on top of Kubernetes is unmatched. It takes out all the complexities of Kubernetes and makes it easier for teams to work even with less/ no knowledge of the underlying infrastructure (Kubernetes). The platform enabled developers to own their work and drastically reduce the dependencies on the devops team for deploying and testing their applications.
Fined-grained Access Control with SSO Integrations: Devtron’s fined-grained access control, it made life easier for DevOps teams. They were able to onboard more developers and create guardrails around the environment where developers could test out their applications without worrying about critical environments. It also made the collaboration a lot easier, strengthening the security posture, and thus empowering developers.
Increased Visibility and Troubleshooting: Devtron’s software delivery platform not only streamlined the CI/CD but also enhanced the Mean Time to Recovery by providing all the necessary information on the dashboard itself making it easier for developers to debug if any issues kick in. Additionally, with Devtron you can observe the application metrics such as CPU and memory utilization, latency, and throughput. Along with that, for the deployed application, they were able to check the logs, and events, and even get into the terminal if any debugging is required which drastically improved the debugging and troubleshooting of applications.
Easy Onboarding on Kubernetes: With features like cloning applications, pipelines, etc, it made it much easier for them to onboard new applications on Kubernetes quickly and hence helped them onboard their applications on GKE in record time.
Conclusion
The company has significantly improved and streamlined the CI/CD pipelines for Kubernetes with the help of Devtron. The months-long target of migrating onto GKE was shortened to weeks, the steep learning curve became a flat line, and the manual tasks are now fully automated. By integrating Devtron, the company has reduced the time-to-market by 30%, increased developer productivity by 40%, and helped them modernize their workloads and provide the scale required for the growth of their business.
Ready to see the same results for your team? Connect with us now!
