ActiveCloud: authorize each instance before failback
A provider’s availability announcement still leaves the decision to restart an instance. The ActiveCloud case supports a failback exercise for situations where contingency services remain active.
Czytaj po portugalskuIn this article
- The notice leaves a decision with the customer
- Name who can keep the service active
- Reconcile work before the transition
- Look for evidence that challenges the authorization
- Prepare a pause and an approved fallback
- Close with the record of a transaction
- Frequently asked questions
- Related reading
- Sources and references
By Pentesty
The notice leaves a decision with the customer
On the public technical channel attributed to the company by The Record’s reporting, the November 11 message says a November 10 ransomware attack encrypted part of its infrastructure. The November 12 notice at 23:48 UTC announced ECP’s return and left VM restarts to customers. The company said it avoided automatic starts because simultaneous operation with alternatives activated at other providers could cause conflicts.
The November 13 summary at 07:59 UTC reported the return of its main services while describing continued assistance for customers experiencing VM problems. The scope here is the November 10–13 communications. Those records establish neither service conditions on November 15 nor that conflicts occurred.
Name who can keep the service active
For Pentesty’s exercise, choose a fictional service and two lab environments: the original and a contingency alternative. Define who owns the operation and can authorize the transition. Record which instance may receive work and which must remain unable to produce effects.
Connect that decision to the scenario’s actual paths: traffic sources, write credentials, queues and schedulers. An address may point to the alternative while an old job remains active in the original environment. Treat that possibility as a hypothesis to test in the lab, without attributing it to ActiveCloud’s architecture.
Reconcile work before the transition
Prepare synthetic transactions with known identifiers. Note which the alternative accepted, which await processing and what result should exist after failback. The service owner should be able to explain how pending work will move, including integrations that perform actions outside the main application.
The Beget recovery analysis examines the association between a backup, account and service. Here, the question is who may continue executing work during the transition. Having a recoverable copy does not, by itself, resolve write authorization across the two instances.
Prepare a pause and an approved fallback
Before execution, describe the symptoms that require a pause: unexpected processing, an incorrect destination or missing evidence about a job. Define who communicates the suspension, who protects pending work and which environment remains the approved operating base while the team investigates.
Plan reversal around effects already produced in the exercise. Routing traffic back does not automatically undo a transaction. The cloud dependency recovery analysis helps map access and capacity needed for the alternative; acceptance of this transition must also track the results of the work performed.
Close with the record of a transaction
Bring together the owner’s signed authorization, instance identifiers, the completed transaction and the rejected attempt. Add pending queues, exceptions and the condition for reviewing the decision. Closure should show who could operate at each stage, without turning the provider’s announcement into automatic acceptance of the application.
Frequently asked questions
Who should authorize the restart?
The organization needs a named service owner who agrees the decision with operations and security. The exercise checks whether that authorization matches the instance that can actually produce effects.
Does the provider’s notice confirm instance conflicts?
The November 12 notice describes a possible risk where customers activated alternatives with other providers. It does not say those conflicts happened or establish an architecture with two active operations.
Does the test require duplicate production services?
The exercise uses two fictional environments, synthetic transactions and restricted destinations. The allowed and rejected conditions are observed in the lab, with boundaries and a stopping condition agreed beforehand.
How should a decision proceed when pending-work records are missing?
Record the gap, who can clarify it and which decision remains suspended. The responsible owner should assess the available conditions without filling the history with assumptions about messages or transactions.
Sources and references
- ActiveCloud: technical ransomware notice, November 11, 2026
- ActiveCloud: VM restart decision, November 12, 2026
- ActiveCloud: recovery milestones, November 13, 2026
- The Record: reporting attributing the channel to ActiveCloud, November 13, 2026
Original Pentesty analysis, published November 15, 2026. The proposed scenarios are editorial guidance, with scope and validation to be defined for each environment.