Cloud Outages Happen – Hybrid Resilience with NDB Keeps Your Databases Running

Why Hybrid Matters More Than Ever

Critical business applications continue to rely on their databases to power everything from customer transactions to internal operations. Moving to the cloud alone cannot be relied on as a guaranteed ticket to delivering disruption-free service, especially given that even hyperscaler platforms occasionally experience outages.

When your mission-critical workloads are hosted entirely on a single platform (either in-the-cloud or on-premises), you could be introducing a single point of failure that can constrain both your operations and recovery options.

This is because failures can cascade across any environment. For example, in cloud-only deployments, full-stack ownership by a single provider can limit visibility and control during outages. A failure in one layer – such as DNS or database services – can ripple across dependent systems, leaving teams unable to intervene directly.

Similarly, on-premises environments are not immune either; localized infrastructure issues can be equally disruptive, especially when recovery paths are rigid or manual.

The path forward isn’t a choice between cloud and on-prem – it’s an embrace of hybrid architectures. Hybrid, in this context, refers to a purposeful combination of on-premises and cloud infrastructure. It’s about designing systems that span environments to reduce risk and improve availability, while simultaneously supporting operational goals like data locality and regulatory autonomy.

To make this practical, organizations must leverage modern hybrid topologies that are resilient by design. That’s where the Nutanix Database Service (NDB) solution stands out. Thanks to its underlying stack, NDB is uniquely positioned to simplify and support hybrid database deployments – bringing together automation, flexibility, and infrastructure awareness in a way that makes business continuity achievable.

This foundation enables organizations to build robust Business Continuity Plans (BCPs) that span both on-premises and cloud environments. By distributing database workloads across these domains, teams can reduce risk and maintain database availability even when parts of the infrastructure go down.

NDB: Your fast track to Hybrid Resilience

Nutanix Database Service (NDB) offers a powerful solution for organizations seeking high database availability with a hybrid design, including for database orchestration and lifecycle management. While you can build a hybrid topology yourself today, NDB can accelerate your journey to Hybrid with an underlying stack that is both cloud and on-prem aware via the Nutanix Cloud Clusters (NC2) solution, which is available on leading public clouds. NC2 avails simultaneous access to clusters across both on-prem and cloud environments, enabling you to leverage these clusters directly in your HA-aware NDB provisioning workflows. In addition to provisioning across hybrid environments, NDB continues to simplify your management overhead across the lifetime of your database via:

• Automated Lifecycle Management for HA databases
• Intuitive GUI, as well as API/CLI-driven automation

This is in addition to availing the Time Machine and Thin Cloning features of NDB to support rapid recovery and testing workflows for your databases.

This combination makes NDB uniquely suited to help businesses achieve true business continuity across diverse environments. One way this comes to life is through resilient cluster topologies that NDB supports out of the box. For example, let’s take a closer look at how NDB can be used to deploy a common hybrid pattern PostgreSQL workload using a 3-Location configuration (within metropolitan area latency) with Patroni to maintain quorum-based availability - even when one location goes offline.

The 3-Metro-Location Topology: A Blueprint for Availability

One of the most effective High Availability configurations designed with a no-data-loss target, is a three-metro-location topology. This setup avoids the common pitfalls that are inevitable with:

• Single Location/Single Cloud deployments, where an outage brings down your entire workload, as well as
• Dual-Location deployments, where losing the primary site (containing 2 nodes of a 3 node HA cluster) can instantly break quorum (majority) and halt operations.

Instead, this 3-location model - built on a low intra-location latency (< 25 ms), over private low-loss network connections, distributes a 3-node PostgreSQL cluster (managed by Patroni) across: 

• Cloud Region (e.g., AWS, Azure) [Primary]¹
• Metro Location A [Sync]
• Metro Location B [Async]

Each PostgreSQL node is co-located with a Distributed Configuration Store solution (etcd in this case) instance, to provide health inputs to Patroni’s quorum management.

NDB can also provision this with built-in support for HAProxy, which works over stretched networks - alternatively, NDB also supports alternative configurations via custom Load-Balancing solutions across isolated networks.

This topology is designed to help maintain quorum and database availability even if one location goes offline, in the event that the other locations remain online and connected.

Figure 1: A 3-node PostgreSQL cluster with Patroni, etcd and a Load Balancer/Proxy distributed across a Cloud Region, and two locations within a Metropolitan location, Metro location A and Metro location B. The load balancers/Proxies and the NDB control plane can be deployed in various configurations - on premises or in the cloud.

What happens during a cloud outage?

In this topology, continuity is achieved because the PostgreSQL cluster spans multiple locations and uses Patroni for quorum-based failover. If the cloud region goes offline, Patroni automatically promotes a remaining node (e.g., in Metro Location A) to primary, allowing operations to continue without manual intervention.

Using NDB on the Nutanix stack dramatically simplifies provisioning HA clusters that span on-premises and cloud environments via the Nutanix Cloud Infrastructure (NCI) and NC2 solutions, respectively. With NDB, teams can define standard configurations and deploy them repeatedly - through an intuitive UI or self-service automation - making resilient topologies practical at scale.

Beyond provisioning, NDB also supports autonomous Time Machines per location. The Time Machine - a crucial capability of NDB - is a built-in data protection engine that continuously captures database snapshots and transaction logs to enable point-in-time recovery (PITR). Each Time Machine operates autonomously per location, maintaining its own recovery history. The snapshots and logs that make up a Time Machine can be stored across multiple Nutanix NCI or NC2 clusters, so that, in the event of a primary-site outage, databases can still be recovered from a remote site. This design provides resilient, cross-site recoverability without relying on external backup tooling, enabling ongoing access to point-in-time recovery features if needed - even during an outage in the primary location. With NDB, these complex hybrid deployments are easy to roll out and maintain - so organizations can focus on continuity, not complexity.

See It in Action

To bring this architecture to life, we’ve prepared a demo that shows how a PostgreSQL cluster that is deployed across a cloud region and two metro locations behaves during a simulated outage.

In the demo, the cluster is provisioned using NDB across:

• Cloud Region – main writable node
• Metro Location A – synchronous replica
• Metro Location B – asynchronous replica

We then simulate a failure in the cloud region and observe how the cluster maintains availability without manual intervention - reinforcing the value of hybrid HA topologies.

Conclusion: Availability without Compromise

Business continuity isn’t just about surviving outages – it's about being able to continue to service your workloads through them. With NDB, organizations can architect and seamlessly roll-out resilient database environments that span cloud and on-prem, while elevating your service levels above what each of them can deliver independently, all without sacrificing simplicity or control. 

Whether you are responding to unexpected disruptions or proactively designing for resilience, NDB empowers you to keep your databases available, your operations running, and your customers satisfied.