As a multi-channel retailer of electronics and technology, Conrad Electronic SE provides many of the items that make our day-to-day life easier, more efficient, and more comfortable. Continuously developing new product concepts for the world of technology, Conrad acts as an inspiration to everyone involved in electronics both professionally or privately.

Procurement: Fast, Simple, and Comprehensive

Conrad’s offerings are available on all channels, the phone, online, or via API, all of which the eCommerce Platform team is responsible for maintaining. To keep their platform cutting edge, and to provide a unique and fast experience for their customers, automation is a core part of their strategy. The platform is a pure microservices architecture, built on top of Java, deployed using Jenkins, and hosted on Google Cloud Platform.

With millions of product updates and requests that are processed in real-time, Conrad needs everything operating cohesively to maintain low-level response times and to stay on top of their SLOs.

“Application performance really matters to us, and most importantly to our B2C and B2B customers. They want to get their shopping done quickly, so Conrad needs to be the fastest and simplest solution available — without errors,” said Tobias Baumgart, Director, E-Commerce Platform at Conrad Electronic SE.

To support these efforts, they’ve begun their journey into implementing Kubernetes.

“We want to harmonize our entire infrastructure within a central Kubernetes cluster, where we offer our teams namespaces within a centrally managed cluster. This provides a solid foundation for operations and helps the team streamline processes,” Tobias added.

Kubernetes = Complexity. Conrad has a plan.

“When it comes to big workloads and the complexity that Kubernetes presents, which is much more than just deploying on Google Compute Engine instances, we are quite confident that we can overcome this. We have a good approach to performance testing, and with GCP, we have a lot of flexibility for bringing Kubernetes into production,” Tobias explained.

Previous to this, with no dedicated Observability solution in place, Conrad’s approach to monitoring was segmented based on certain components such as CPU, RAM, or JVM metrics. But after implementing DevOps practices and migrating to a microservices-based architecture, Conrad Electronic realized the need for a fully integrated Observability platform that served as a single source of truth.

“Because our teams and strategies were changing, we needed a solution that brought all the pieces together, creating transparency for a successful digital transformation,” Tobias noted.

When Conrad migrated from on-premise to GCP, Instana provided insights into how the applications and services behaved. From deploying Instana’s agent into Google App Engine to utilizing Google Cloud Functions, instrumenting Instana in Google services provided a comprehensive view of their entire platform.

Transparency Enables Responsibility

To reduce SLO breaches, the eCommerce team creates custom SLI and SLO dashboards that display and analyze the performance of services over time.

“Instana’s one of the core tools that manage our operations. All of our metrics, SLOs, and SLIs are defined within Instana and triggers our alerting chain. It’s the backbone of our SRE team and the whole platform operations,” added Tobias.

With Instana in place, Conrad has complete visibility and understanding of their containerized microservices in GCP, reducing the amount of time previously required to monitor, pinpoint, troubleshoot, and resolve application performance issues.

“We have more transparency, which gives the development teams more responsibility. As we drive the platform with more and more confidence, we are seeing that our systems are working and that the teams are fixing issues that they’re are responsible for,” mentioned Tobias.

This transparency enabled Conrad to introduce a post mortem culture, which helps them overcome production issues, improve team communication, and reduce finger-pointing. Overall, the success and productivity of the teams have increased, which has a positive impact on resolving incidents quickly.