JOURNEY 02 · RESILIENCY & RECOVERY

Recover environments,not just data.

FluidCloud backs up your entire cloud in snapshots: every network, role and dependency, which you can restore in minutes.

SOLUTIONS / CLOUD RESILIENCE & DISASTER RECOVERY

Cloud resilience isan infrastructurechallenge.

Backups restore your data. FluidCloud clones everything underneath it: every VPC, IAM role and route. You can recover a region, an account or a full cloud.

02 / THE 2025 RECORD

Last year, the cloudwent down. A lot.

On October 20, 2025, a failure in the AWS load balancer monitoring broke DNS resolution for DynamoDB in US-EAST-1. Within an hour, more than a thousand services were degraded. They stayed degraded for about fifteen hours.

GCP, Cloudflare and Azure also had major incidents that year. Each one failed in the same layer: the infrastructure under the workloads.

  1. DORA designationAWS, Azure, and GCP named critical ICT providers
  2. AWS US-EAST-1DNS · about 15 hours
  3. Azure StorageRegion unavailable
  4. CloudflareDNS · CDN edge
  5. GCPAuth · IAM
  6. Azure ADIdentity · IAM
03 / CLOUD CLONING™

Discover. Map. Clone.

One mechanism for all three: recovery after an outage, a rebuild after a breach, or a move to another cloud.

  1. STEP 01

    Discover

    Connect with read-only access. FluidCloud reads every VPC, IAM role, security group, secret and dependency through the provider APIs. There are no agents and nothing to install.

  2. STEP 02

    Map

    Each resource becomes a node, and each dependency becomes a typed edge. Each node has its tags, usage, cost and policy. The result is one versioned graph that stays current.

  3. STEP 03

    Clone

    Make the topology again in another region, account or cloud. Or restore it in place to an earlier known-good state. You get a running environment, not a document.

  • AWS
  • Azure
  • GCP
  • OCI
  • VMware
  • Nutanix
  • Vultr
  • OVHcloud

Discovery is read-only. A clone writes only to the account that you give it access to.

04 / ONE MECHANISM, THREE FAILURE DOMAINS

Recover across a region, anaccount, or a whole cloud.

The fastest DR pattern, with the same topology as the primary.

Clone the topology from the latest capture into another region. Test it, then switch over when it passes. FluidCloud never changes the source. Recovery time depends on the resource types and the data size.

Typical failover
~4m 12s
Topology
From the latest capture
Standby idle cost
−68%
A FOURTH PATTERN · EVENT-DRIVEN

Resilience without paying to keep a second copy warm.

Capture the topology on a schedule. Run nothing idle. Regenerate the topology when you need it. A person, a pipeline or an agent can start the recovery from the console, the API or MCP.

  1. 01 · SIGNALYour alert or scheduleDatadog · CloudWatch · cron
  2. 02 · TRIGGERYour alerting calls the APIOr an agent, through MCP
  3. 03 · CAPTUREOn-demand scanRead-only · versioned graph
  4. 04 · DEPLOYRegenerate the topologyOnly when you need it
05 / RISKS, MAPPED

Other failures. One operation.

These failures use the same recovery operation. None of them needs its own runbook.

06 / BUILT FOR THE AUDITOR

Every recovery produces its ownpaperwork.

There is no “export the runbook” step. The operation itself makes the record of what ran.

  1. BEFORE

    Assessment report

    Every resource and every flagged gap. You review it before you deploy.

  2. DURING

    Apply log

    Every change is logged as it runs. If a run fails partway, what was built stays. Fix the cause, then apply again to continue.

  3. AFTER

    Kept record

    You keep the Terraform, the state file and the full log of every run.

  4. DRILLS

    Rehearsed recovery

    A drill restores a capture into an account that you choose, never into production.

DORA · ARTICLES 28–31Regulated firms must show a tested exit from critical providers. A cross-cloud clone is that exit.

  • SOC 2 Type II
  • GDPR
  • Read-only access
  • No agents
07 / WHERE DO YOU STAND?

How regenerable is yourinfrastructure today?

RESILIENCE SELF-ASSESSMENT

Five questions. Under two minutes.

It measures capability, not intent. It shows the gap between your DR plan and what would happen if a production account failed tonight.

  1. 01Hopeful
  2. 02Partial
  3. 03Operational
  4. 04Regulated
Start the assessment

DOWNTIME-COST CALCULATOR

What is an hour of downtime worth to you?

Hours avoided / yr
8.8h
Annual avoided cost
$0.42M

These are estimates from modeled assumptions. They are not a guarantee.

08 / IN PRODUCTION

Customer voice. Industryvalidation.

“Our last cross-cloud migration took eleven months. The next one took a Tuesday afternoon.”
VP Platform Engineering · regulated healthcare
“This isn’t just a migration tool. The platform naturally supports rollback, inventory, and multi-cloud visibility use cases.”
Lead Architect · enterprise security company
90%
Reduction in migration effortVultr · production
10k+
Resources clonedEnterprise cloud migration
<1 week
7-year AWS estate made resilientFinancial services
09 / FAQ

Questions resilienceteams ask.

Does FluidCloud replace our backup tools?

No. Backup tools protect your data. FluidCloud protects the infrastructure topology under the data: VPCs, IAM, routing, security groups and dependencies. You restore the data into an environment that FluidCloud has already rebuilt.

Is FluidCloud a Terraform or IaC generator?

No. FluidCloud keeps one versioned graph of your estate: each resource is a node and each dependency is a typed edge. Terraform is one output of that graph. You review it before you run it.

Do we need to run active-active?

No. A warm standby means that you pay for idle topology. Capture the topology on a schedule, run nothing idle, and clone it from the latest capture when you need it.

Does FluidCloud need agents or write access?

Not for discovery. Discovery reads through the provider APIs with read-only access. There are no agents and nothing to install. Recovery builds in the target account, so it uses write permissions that you scope and approve for that account.

Can we test recovery without touching production?

Yes. Restore any capture into a separate account or region to practice. Before it deploys, you get an assessment report of every resource and every flagged gap.

Which clouds are supported?

AWS, Azure, GCP, OCI, VMware, Nutanix, Vultr and OVHcloud.

10 / YOUR MOVE

Point us at one account.

Map one cloud account with read-only access. You keep the topology graph and the recovery plan, even if you never talk to us again.