ABAX tracks over 500,000 assets with entirely cloud native strategy | SUSE Communities

ABAX processes a vast amount of data from over half a million devices every hour, making uninterrupted availability critical to the company's success. To ensure reliability, the company turned to Kubernetes as its primary compute platform, migrating numerous IT processes to the container orchestration system. ABAX runs most of its Kubernetes clusters on Google Public Cloud, with some operations on AWS and Microsoft Azure.

As a proponent of open source, ABAX relies on several open source solutions, including Rancher Prime, K3s and Traefik Enterprise, to optimize its infrastructure operations and deliver services to clients. These tools allow ABAX to manage and deploy its Kubernetes infrastructure efficiently and securely, maintain granular control over user access, and provide next-generation customer services.

“Rancher Prime and Traefik Enterprise have become an integral part of the ABAX cloud native infrastructure,” says Ornell. “With Traefik Enterprise, everything is automated and easy to use. We maintain functionalities that our developers rarely see because they just work.”

Regardless of its dedication to open source, ABAX understands the importance of risk mitigation and utilizes enterprise support to ensure its systems remain stable and secure. By partnering with SUSE and Traefik Labs, ABAX can focus on building its future and delivering exceptional services to its clients with peace of mind.

Maximizing productivity with a consistent experience

ABAX, with a lean IT team handling various platform stacks and tools, aims to provide a consistent experience, performance and troubleshooting to its application developers. To achieve this, the IT team needs complete control over the platform stack. Rancher Prime and Traefik Enterprise make ABAX’s vision of a cloud native, infrastructure-agnostic platform a reality, providing developers with a single, consistent experience, regardless of the underlying cloud infrastructure.

The Rancher Prime interface allows developers and operations to work together in production with ease. The experience is the same, no matter which cloud is hosting a cluster. “We don’t have to train our developers how to use the different user interfaces between the different cloud providers,” says Ornell. “Rancher Prime’s unified control system is key to maximizing productivity.”

Saving time with automated RBAC

Rancher Prime’s automated role-based access controls (RBAC) have also made a significant impact in enhancing the IT team’s productivity. Automated RBAC enable the team to grant and limit access to clusters and defined projects within Rancher Prime easily. If a developer works on a particular business domain, these RBACs can grant access to only that business domain inside the production cluster, which is particularly important in highly regulated customer industries like insurance.

“What we're doing is giving developers direct access so they can manage their own workloads to a certain extent, saving us valuable time,” says Ornell. “This is key for us to be able to survive as such a small team managing so many different things.”

Simplifying Kubernetes deployments with Fleet

Fleet, a feature of Rancher Prime, simplifies managing, deploying and scaling containerized applications across multiple clusters. It simplifies the process of managing distributed application deployments, allowing users to easily start, scale and keep track of the containers that make up their applications. It also saves time and resources by automating how users set up applications in different environments.

Fleet currently runs 51 infrastructure deployments and maintains more than 70 Kubernetes clusters for developers, which would be impossible for the IT team to manually manage otherwise. “We label our clusters with the features we want them to have,” says Ornell. “For example, we’ll label a cluster with Traefik Enterprise. Fleet then automatically deploys Traefik Enterprise to that cluster. Same with Datadog or any other infrastructure piece we might need in a cluster. This is key to being able to do a disaster recovery swap quickly because we wouldn’t have time to manually reconfigure everything when in a disaster scenario.”

Faster disaster recovery with a powerful trifecta

Speaking of disaster scenarios, ABAX was able to achieve an impressive disaster recovery time for its entire production compute environment (excluding the database environment) in just one hour 15 minutes, thanks to three features of Rancher Prime.

Rancher Prime's multi-cluster management enabled the team to efficiently monitor and control resources across multiple clusters through a single pane of glass, while RBAC and Fleet made deploying fine-grained control over user access to resources across multiple clusters significantly easier. With these three components, the team was able to relaunch its operations efficiently and securely at scale, saving precious time.

According to Ornell, “Getting the environment itself back online wouldn’t take that much longer without Rancher Prime. The real difference lies in the post-recovery work, such as providing proper access for developers, ensuring all infrastructure components are back in the cluster and so on. Rancher’s RBAC and Fleet make these steps significantly faster. For comparison, doing this manually would take days.”

Standardizing cluster environments with K3s and Traefik

ABAX uses K3s to standardize cluster environments between development and production. K3s is a lightweight Kubernetes distribution originally developed by Rancher Labs and is now a CNCF project. K3s uses Traefik as the default ingress controller to manage incoming traffic. “The tiny differences between developer machines and production environments leave room for error, especially when trying to figure out why a service is working on a developer machine but not on a production environment,” says Ornell.

The team implemented k3d, a lightweight wrapper used for running K3s in Docker, to run small Kubernetes clusters on every developer machine. “K3s, deployed through k3d, enables us to deploy everything in a developer environment as if it were a production environment, making it easier to identify issues when they arise,” says Ornell.

Consequently, nothing has to change in how the Kubernetes and networking layers operate when applications move from development to production, speeding time to launch.

Resolving issues with SUSE and Traefik Labs’ enterprise support

SUSE’s commitment to open source ethos means that Rancher technology is the same whether it comes from the community or enterprise version. For resource-constrained teams, paid support can be invaluable.

In working with SUSE Support, Ornell has found the team resolves issues before other vendors, working on the same issue, deliver a second reply. “SUSE has the competency to debug Kubernetes comprehensively, so we can figure out what is actually going wrong,” says Ornell. “We have very limited resources available. If we can offload debugging and figuring out what’s wrong to a support team instead of us having to do it, that’s golden on our end”

Support goes to another level when support teams from SUSE and Traefik Labs collaborate to resolve issues. Reflecting on a particular instance, Ornell states, “We couldn’t pinpoint where the issue was and we ended up being on a support call with support representatives on both sides collaborating, which to me is invaluable.”

Accelerating solutions via a partner ecosystem

SUSE’s partner network also provides additional value for those working with the company. This is how ABAX, for instance, discovered Traefik Labs when it sought to employ a Kubernetes-native ingress controller to manage incoming data from various client machinery, assets and automobiles. After browsing SUSE’s partner application catalog for Rancher Prime, ABAX selected Traefik and immediately integrated it into its infrastructure. Now, Traefik Enterprise handles every incoming request. Traefik Labs also played an essential role in ABAX’s cloud migration acceleration. Moreover, all customer-facing portals run on Traefik via a GKE cluster in Google Cloud, provisioned by Rancher Prime.

What’s next?

Looking ahead, ABAX, now a major player in the European telematics market, plans to leverage the data it already processes to provide additional customer value. By embracing data sciences and artificial intelligence, ABAX customers will soon gain new insights that will inevitably lead to greater innovations. Backed by open source solutions and expertise, ABAX has a bright future ahead.

This is a companion discussion topic for the original entry at https://traefik.io/blog/abax-cloud-native-strategy/