Introducing Cloud 66 Deploy v3 — A Next‑Generation Kubernetes Deployment Platform 🎉
We’re excited to announce the rollout of Cloud 66 Deploy v3 — a major evolution of our deployment platform built from the ground up for flexibility, modern workloads, and future growth.
While Deploy v2 has served many teams well over the years, Deploy v3 brings a fresh architecture and new capabilities designed for today’s cloud-native needs. Here’s what you need to know.
Why Deploy v3 — What’s Changed
Deploy v3 is not just an incremental upgrade — it’s a reimagining of how deployments should work in a Kubernetes world. Under the hood, both v2 and v3 share core components (Kubernetes, BuildGrid for image builds, and our hosted image registry), but v3 rethinks how the pieces fit together.
At‑a‑Glance Comparison
| Feature Area | Deploy v2 | Deploy v3 |
|---|---|---|
| Auxiliary services | Typically external/managed | Run stateful/aux services (e.g., databases) inside the cluster |
| Availability modes | No built‑in HA tiering | RA (Reduced Availability) and HA (High Availability); RA → HA upgrade path (planned) |
| Cloud provider support (initial) | Broad (AWS, Azure, GCP, DO, etc.) | Hetzner and Vultr to start |
| Cluster model | Server‑centric; manual scaling and role assignment | Kubernetes‑native; decoupled architecture with RA/HA cluster types |
| Manifest handling & extensibility | Directly speaks to Kubernetes API | First‑class Kubernetes manifests generated and customizable |
| OS & infrastructure | Ubuntu (mutable) | Flatcar Linux (immutable/read‑only) |
| Persistent storage | Per‑server or external | Shared persistent storage via CSI (e.g., Longhorn, cloud disks) |
| Supported hardware | Mostly x86 | x86 and ARM; GPU via node pools |
| Workload flexibility | Single node roles; limited specialization | Node pools (e.g., GPU, high I/O, ARM/x86) to target workloads |
Notes: CSI is storage; CNI (networking) is supported by Kubernetes but not listed here as a storage feature. RA → HA upgrades are planned; HA → RA is not supported.
Cluster Types: RA vs HA — Choose Based on Your Needs
Reduced Availability (RA). A lower‑cost, lower‑redundancy mode. Ideal for development, testing, staging, or smaller production workloads. A lighter footprint with all of v3’s architectural benefits. Planned upgrade path to HA without re‑architecting.
High Availability (HA). Built for resilience and mission‑critical workloads. Redundancy across nodes and stronger reliability guarantees. Use HA when uptime, fault tolerance, and production‑grade durability are essential.
The Big Benefits You’ll See
- Flexibility & scalability — Start small with RA, then upgrade to HA when needed.
- Resilience out of the box — HA clusters are designed for redundancy and uptime.
- Kubernetes‑native architecture — No more tight coupling to servers; manifests are exposed and customizable.
- Modern workload support — Run GPUs, ARM/x86 nodes, and stateful services inside the cluster.
- Better storage & I/O handling — Shared persistent storage with cloud‑native support under the hood.
- Improved security & maintenance — Flatcar Linux (read‑only) reduces drift and patching overhead.
- Future‑readiness — v3 will rapidly expand cloud coverage and close feature gaps while introducing new capabilities.
Rollout, Support, and Migration
- Early Access. Deploy v3 is available in early access and free to try while we collect feedback from early adopters.
- Parallel support. Deploy v2 remains fully supported. As v3 matures, we aim for feature parity and beyond without destabilizing existing workloads.
- Pragmatic migration. Keep stable, mature workloads on v2. Use v3 for new projects or dev/staging to learn and shape the future path. Migrate as v3 fills the gaps you need.
When to Use v2 vs v3
- Choose v2 if your workload depends on mature integrations, broad cloud support, or features not yet in v3. If stability is paramount today, v2 remains the safest bet.
- Choose v3 to embrace Kubernetes‑native workflows, node pools, modern hardware (ARM/GPU), and full manifest customization — and if you’re happy to be an early adopter.
- Use both concurrently: keep mission‑critical production on v2 while exploring v3 for new environments.
Deploy v3 is a bold step forward: flexible, extensible, and built for the evolving needs of cloud‑native applications. We invite you to try Deploy v3, explore its capabilities, and help us shape the roadmap.
For a deeper dive into differences and guidance, see our documentation.