LinkedIn Replaces Argo CD with Custom GitOps Engine for Enhanced Kubernetes Deployment Scalability and Reliability
LinkedIn has announced that it has replaced Argo CD with its own custom-built GitOps engine for managing Kubernetes deployments. This new system is responsible for deploying applications across more than 250 Kubernetes clusters, supporting over 4,000 services, and handling upwards of 100,000 deployments weekly. The migration, which occurred between July 2025 and April 2026, has resulted in a dramatic reduction in deployment infrastructure failures, from 925 to 63, despite a 3.7x increase in deployment volume. Furthermore, the median time to detect and resolve failed deployments has fallen from 16.1 minutes to just 2.45 minutes.
This development is significant for organizations operating at scale, particularly those pushing the boundaries of cloud-native infrastructure. It demonstrates that even leading GitOps tools like Argo CD, while excellent for many use cases, may not always meet the unique demands of hyper-scale environments. LinkedIn's decision to build a custom solution was driven by specific challenges encountered with Argo CD at their scale, such as issues with linking sync and health status to specific deployments due to the lack of an `observedGeneration` field, and problems with stale repo-server cache entries leading to unintended revisions. This move underscores a broader trend where large enterprises are increasingly willing to invest in custom engineering to overcome the limitations of generic tools when operational efficiency and reliability become paramount.
This trend aligns with the ongoing evolution of GitOps, which emphasizes declarative infrastructure and application management through Git. While tools like Argo CD and Flux have become de facto standards for implementing GitOps, LinkedIn's experience suggests that for organizations with extreme scale and complexity, a deeper level of customization may be necessary. Other large operators, such as Uber with its Up platform, have also adopted similar approaches, building deployment logic around Kubernetes rather than relying solely on general-purpose GitOps controllers. This indicates a maturation of the GitOps landscape, where the core principles remain, but the implementation details become highly specialized to meet specific enterprise needs.
In practice, this means that while most organizations will continue to benefit immensely from established GitOps tools, those operating at LinkedIn's scale should be prepared to evaluate whether off-the-shelf solutions can truly meet their long-term scalability and reliability requirements. Practitioners should closely examine the specific pain points LinkedIn identified, such as the need for per-deployment isolation and robust health checking mechanisms, as these are common challenges in large-scale deployments. It also highlights the importance of contributing to open-source projects or, when necessary, developing internal solutions that can be precisely tailored to an organization's unique operational DNA. The trade-off involves significant engineering investment versus the potential for unparalleled control and performance gains.
Read original source