Google Cloud previews in-place migration to GKE Dataplane V2
Google Cloud has introduced a preview of in-place migration for eligible VPC-native clusters moving from GKE Dataplane V1 to Dataplane V2, its eBPF and Cilium-based networking layer for Google Kubernetes Engine.

Key takeaways
- Google Cloud has launched a preview of in-place migration from GKE Dataplane V1 to Dataplane V2.
- Eligible VPC-native clusters running GKE 1.35 or later can migrate without a parallel cluster setup.
- Routes-based clusters and Calico clusters are not supported during the preview.
- Google Cloud advises testing on non-production clusters first.
In this article
What changed
Google Cloud announced the preview of in-place migration to GKE Dataplane V2 on 19 November 2026, according to the Google Cloud Blog. The feature lets eligible VPC-native clusters move from GKE Dataplane V1 to Dataplane V2 without setting up a separate parallel cluster. Previously, customers often faced a disruptive migration process when moving production clusters with stateful workloads and existing IP allocations.
How the migration works
The process runs in three stages. First, administrators run eligibility and preflight checks, including verifying that Pod IP ranges have enough capacity to recreate nodes and checking for incompatible configurations such as custom eBPF programs on managed interfaces. Second, the migration starts with a single Google Cloud CLI update command. Third, GKE updates node pools to Dataplane V2 while respecting maintenance windows, exclusions, and upgrade strategies. Customers can patch specific node pools manually to move faster, or roll back to Dataplane V1 at any time before finalization.
Requirements and limits
In-place migration supports VPC-native clusters running GKE version 1.35 or later. Routes-based clusters and clusters running Calico are not supported. During the preview period, Google Cloud recommends testing migrations on non-production or test clusters before planning a broader transition.
What Dataplane V2 offers
GKE Dataplane V2 is built on extended Berkeley Packet Filter, a Linux kernel technology for packet processing, and open-source Cilium. It replaces traditional iptables routing and kube-proxy. Google Cloud says it supports large clusters with network policy enforcement and handles high Pod churn and fast sandbox provisioning for large agentic AI workloads. It also serves as the networking foundation for GPU and Cloud TPU clusters on GKE and powers GKE managed DRANET for Dynamic Resource Allocation of network devices.
Security and observability

Dataplane V2 enforces standard Kubernetes NetworkPolicy and hierarchical ClusterNetworkPolicy directly in the kernel with optimized performance. It provides the underlying dataplane for sidecarless GKE ambient networking. Observability features include native network policy logging, live traffic topology, and Hubble flow metrics.
Sources
- Introducing in-place migration to GKE Dataplane V2Google Cloud Blog
Spotted a mistake? See our corrections policy. Read how News by OnDevice works in our editorial policy.



