How CommonGround Boosted R&D Operations with groundcover

Introduction

Based in Silicon Valley and Tel Aviv, CommonGround aims to improve communications and online collaboration. Founded in 2019 by serial entrepreneurs, the company is led by a world-class team of engineers with expertise in AI, computer vision, geometry processing, 3D graphics, and real-time video compression.

The problem

CommonGround's DevOps team was using New Relic for its Kubernetes clusters that was causing problems due to its high hardware footprint, which was suffocating their worker nodes with too many requests for CPU and memory. Additionally, the user interface was far from user-friendly, making it difficult to understand. It also didn’t provide valuable information for debugging."We projected the user interface on our screens in the operations room, and no one seemed to know what was going on at any given time. You looked at the monitor and it told you nothing." Ido Ozeri, DevOps & MLOps Team Leader at CommonGround.

{{quote-block}}

Why groundcover?

Ido led the evaluation of multiple alternatives. In the end, CommonGround’s decision to move forward with groundcover was influenced by multiple factors:

• Product maturity - this was a key factor for Ido and the team. The availability of many different extensions within groundcover enabled a high degree of configurability, in addition to the ability to control and customize their data, were hugely advantageous, and were not possible with other solutions. The use of open-source projects was another important factor.

• Ease of implementation - Ido’s team started using groundcover on just one cluster. They could not believe how easy the deployment was and how quickly they were able to show data on all their applications. Within minutes, it was all there for them to start exploring.

• Customer support - immediately after deploying, the DevOps team joined groundcover’s Slack community. This brought them huge value by being able to communicate directly with groundcover’s experts, who were able to offer tips and best practices that helped them be even more efficient in their ability to identify and solve issues.

• Budget planning ability - groundcover’s pricing scheme, which is based on the number of nodes, enabled a fully predictable expenses forecast. This helped make higher management much more comfortable approving the budget, which was also considerably lower than what their other options would have required.

The value

Since the implementation, groundcover has expanded its user base at CommonGround, now reaching the entirety of their R&D department. This includes two QA engineers, four backend developers and four DevOps engineers, which comprises four engineers. For the DevOps team, groundcover is a core platform for monitoring, investigations and response. QA engineers completely transitioned to using groundcover as their primary log analytics tool, instead of what they had been using. The backend team has made it the first place they go to to investigate all issues they identify. They are also leveraging the custom metrics and alerts capabilities.

The Impact

The impact of using groundcover was immediate. From day one, they were able to detect issues that they didn't even know they had. The system was able to present and summarize issues in a straightforward and succinct manner, making it much easier to detect errors. For example, CommonGround run backend services written in Go on AWS and machine learning pipelines written in Python on GCP. They were experiencing many failures that were not necessarily related to the code but rather infrastructure failures. These failures were difficult to detect as logs were coming from multiple sources. With groundcover, they were able to instantly see the root cause of every pipeline failing, which was a major time saver.