Blog de Zscaler
Reciba en su bandeja de entrada las últimas actualizaciones del blog de Zscaler
Go Dark or Go Unprotected? Why Your Workloads Need a Better Business Continuity Plan
A global bank processing transactions, a large healthcare provider syncing patient systems, or a manufacturer running real-time plant operations can't tolerate unplanned downtime or unsecured fallback traffic. For these production workloads — secured by a cloud-native zero trust architecture — an outage isn't just an inconvenience. It can mean lost or unsecured connectivity for systems that were never designed to tolerate downtime.
The Zscaler Zero Trust Exchange™ is the world’s largest security cloud, protecting 750+ billion transactions daily. With a globally distributed network of more than 200 public data centers across 185 countries, it seamlessly absorbs component, network, and regional failures to deliver secure, low-latency access to companies worldwide. Yet, no system is immune to unforeseen events — which is why you need a business continuity plan to keep workloads connected and secure through catastrophic failures.
3 Pillars of an Efficient Disaster Recovery
An efficient disaster recovery plan should ensure the following:
Stay connected: When the Zscaler Public Service Edge can’t be reached, workload traffic moves to the private service edge - a predefined backup cloud (Zscaler managed Business Continuity Cloud or customer managed data center).
Stay inspected: The workload traffic is still inspected by Zscaler private service edge based on the policies defined in the Public Service Edge.
Come back on its own: When the Zscaler Public Service Edge is healthy, workload traffic moves back from the private service edge automatically.
Introducing Business Continuity Cloud for Workload Communications
Business Continuity Cloud — Zscaler's existing business continuity and disaster-recovery capability for users — is now available for workloads. It adds a private data plane and control plane that protects workloads traffic with uninterrupted security, even when the public cloud is unavailable or unreachable. Here's how it works.
Managing the Data Plane Outage: Automated Failover
Typically, when a workload tries to connect to the internet or a SaaS destination, traffic is sent to the Cloud Connector, which forwards it to the Zscaler Zero Trust Exchange (ZTE). Customers inspect this traffic and enforce advanced threat and data protection using Zscaler Internet Access (ZIA). If the Zscaler Public Service Edge (ZTE) goes down, egress traffic is either blocked entirely (fail-close) or allowed through without inspection (fail-open).
With Business Continuity Cloud support for workloads, traffic instead fails over automatically to a dedicated, geographically separate standby cloud instance that enforces the same policies defined in the Zscaler Public Service Edge. Customers can use the Zscaler-managed Business Continuity Cloud (BCC) or deploy a Zscaler Virtual/Private Service Edge in their own data centers to handle the inspection.

Figure: Traffic going through Zscaler managed BCC during data plane failure

Figure: Traffic going through customer managed Data Center during data plane failure
Managing the Control Plane Outage: DNS triggered failover
What do you do when the Zscaler admin portal itself is unreachable? You can't make changes, flip a switch, or respond — right when you need to most. For these situations, customers can pre-configure a DNS TXT record that Cloud Connector polls periodically, giving them an external "switch" to trigger failover.
This also makes fire drills practical, not just emergencies — customers can validate failover on a defined subset of traffic on a regular cadence, without redirecting all production traffic, so continuity plans are proven before they're ever needed.

Figure: DNS triggered business continuity during control plane failure
Benefits
Together, automated failover and DNS-triggered failover give workloads the same resilience that's long protected user traffic — without asking customers to choose between staying online and staying secure. Here's what that translates to in practice.
Production continuity independent of security-layer availability
Until now, the health of your production workloads was tied to the health of a single cloud control point — if the Zscaler Public Service Edge went down, so did your options. With business continuity built in, that dependency breaks. Workload traffic keeps flowing to its internet and SaaS destinations through the Zscaler Business Continuity Cloud (BCC) or customer managed Data Center with Zscaler virtual/private service edge, fully inspected, regardless of what's happening upstream. Production uptime is no longer hostage to the availability of any one component in the chain.
No forced trade-off between staying secure and staying online
Fail-open and fail-close have always asked customers to choose one: let traffic through unprotected, or cut it off entirely. Neither is a real continuity plan — just damage control. Automated failover to the Business Continuity Cloud or a customer-managed private service edge removes that choice altogether. The same policies enforced at the Zscaler Public Service Edge carry over to the backup path, so traffic stays both online and inspected, even during an outage.
Low operational overhead
Resilience shouldn't require a second team to run it. Failover to the private service edge is automatic — no manual cutover, no scripts to trigger in the middle of an incident. Provisioning happens through existing licensing rather than a separate deployment project, and configuration lives in one place rather than being repeated across every gateway. Testing is self-service too: customers can run fire drills on a defined slice of traffic whenever they want proof the failover path works, without waiting on Zscaler or disrupting production to find out.
Fits enterprises with strict uptime and compliance requirements
For regulated industries and large enterprises, disaster recovery isn't optional — it's something auditors and boards expect documented and tested for every critical vendor in the stack. Business continuity for workloads gives these organizations exactly that: a defined failover path, a way to validate it on a schedule, and a control mechanism that works even if the admin portal itself is unreachable. It's DR that can actually be demonstrated, not just assumed.
Conclusion
Outages are no longer a question of if, but when — and for critical production workloads, "when" needs an answer that doesn't force a choice between staying online and staying secure. By extending Business Continuity Cloud from users to workloads, Zscaler closes that gap on both fronts: automated failover keeps traffic inspected and flowing when Zscaler Public Service Edge is unreachable, and DNS-triggered failover gives customers a way to act even when the admin portal itself is down. This isn't just a fallback — it's a business continuity plan enterprises can actually validate before they ever need to rely on it. Zero Trust and business continuity were never meant to be separate conversations, and now, for workloads, they aren't.
To learn more about how Zscaler secures your workloads, click here.
¿Este post ha sido útil?
Exención de responsabilidad: Este blog post ha sido creado por Zscaler con fines informativos exclusivamente y se ofrece "como es" sin ninguna garantía de precisión, integridad o fiabilidad. Zscaler no asume ninguna responsabilidad por errores u omisiones ni por las acciones que se tomen basándose en la información proporcionada. Cualquier sitio web o recurso de terceros enlazado en esta publicación de blog se proporciona únicamente por conveniencia, y Zscaler no se hace responsable de su contenido ni de sus prácticas. Todo el contenido está sujeto a cambios sin previo aviso. Al acceder a este blog, acepta estos términos y reconoce ser el único responsable de verificar y utilizar la información de manera adecuada según sus necesidades.
Reciba en su bandeja de entrada las últimas actualizaciones del blog de Zscaler
Al enviar el formulario, acepta nuestra política de privacidad.


