The mission must continue
- Uptime SLA
99.999%
- to healthy locations in the Oct 2025 AWS us-east-1 outage
- 10 s
- for a failing replica to leave the pool, by default
- ~30 s
What Control Plane keeps running
An alternate site that is already serving
In a Global Virtual Cloud the alternate site is already serving. Health checks and geo-DNS move users to it when a location fails. With the defaults, a failing replica leaves the pool in about 30 seconds: one check every 10 seconds, three failed checks.
Control Plane provides the running alternate site for your applications. People, essential functions, facilities, succession, communications and data replication stay part of your COOP plan.
VMs · Bare Metal · Edge — Networking · Identity · TLS · Observability · Failover
Your data centers and clouds as one set of locations
Servers in your own data center join Control Plane as a location through Managed Kubernetes, and run the same workloads as any cloud region. A cloud location can back up your site, or your site can back up a cloud.
Four providers usually mean four consoles, four identity systems and four sets of runbooks. Here they share one network, one identity model and one audit trail.
Runs in your own accounts · Credits and commitments still count
Applications that stay inside the locations you approve
A Global Virtual Cloud lists exactly which locations a workload may use, and the same spec deploys to any of them. Moving an application to another approved location is a change to that list, not a rebuild.
Least-privilege access · Micro-segmentation · Encryption · Immutable audit trail
Continuity you can test in an exercise
Take a location out of routing during a continuity exercise and measure what users saw: request success, latency, and how long they took to land on the next location. Every change made during the exercise, by people or by AI agents, is recorded in the same audit trail.
Write down exactly what was disabled. Draining one region's replicas proves something different from losing a whole cloud provider.
How essential applications stay up
Example: your data center with cloud failover
An organization runs an essential application in its own data center and needs it to keep running if that site goes dark.
Join the data center
The data center's servers join Control Plane as a location through Managed Kubernetes, so they run the same workloads as any cloud region.Add two providers
Locations on two different cloud providers join the same Global Virtual Cloud.Set the failover order
The data center serves first. The cloud locations sit on a standby routing tier and take traffic only when every location ahead of them is down.Test it in a continuity exercise
Take the data center out of routing during an exercise and log how long users took to land on the cloud locations.
cpln apply --file essential-app-gvc.yaml
kind: gvc
name: essential-app
spec:
staticPlacement:
locationLinks:
- //location/my-dc-east # your servers
- //location/aws-us-west-2
- //location/azure-eastus2
locationOptions:
- locationLink: //location/my-dc-east
routingTier: 0
- locationLink: //location/aws-us-west-2
routingTier: 1 # standby
- locationLink: //location/azure-eastus2
routingTier: 1 # standby
Security and compliance
The platform's security model is the same on every provider and in every data center you connect.
Zero-trust networking
Deny-by-default firewalls on every workload, and mutual TLS between workloads across regions and clouds.Immutable audit trail
Every action by people, service accounts and agents is recorded and attributable.SOC 2 Type II
Control Plane completed the AICPA SOC 2 Type II audit, performed by Prescient Assurance.
FAQ
The infrastructure under your essential applications: where they run, how users reach them, and how they keep serving when a site, a region or a cloud provider fails.
A complete continuity plan also covers people, essential functions, facilities, succession, communications and records. Those stay with your organization.
Health checks stop sending it traffic and geo-DNS routes users to the locations still serving. With the defaults, a failing replica leaves the pool in about 30 seconds: one check every 10 seconds, three failed checks.
Locations on a standby routing tier take traffic only when every location ahead of them is down.
Yes. Servers in your data center join as a location through Managed Kubernetes and run the same workloads as a cloud region, in the same Global Virtual Cloud as your AWS, GCP, Azure or OCI locations.
Workloads with at least two replicas in at least two locations are backed by a 99.999% SLA.
During the October 2025 AWS us-east-1 outage, 20% of Control Plane customers had us-east-1 in their Global Virtual Clouds. Their workloads were active and running in healthy locations within 10 seconds, with no customer downtime, according to Control Plane internal incident telemetry.
Test your continuity before you need it
Book a continuity architecture review. We'll map one essential application and its dependencies, then show a location dropping out of routing while the app keeps serving.

