PENTESTY · EDITORIAL ANALYSIS · · 4 min read
Columbia: who tracks requests when the portal goes down?
An alternative phone route needs to lead to traceable service. Check how a request gets an owner, avoids duplication and returns to the main queue.
By Pentesty
The notice describes partial impact at a defined time
Cyber Security Brazil’s report prompted this analysis. In its October 21, 2026 notice, Columbia reported an incident affecting MyColaSC and other services. It said public safety, water and the Payment Center, including online payments, remained operational, with SLED leading the investigation. It provided a phone route for non-emergency requests. This is the disclosed account, not a current availability measurement.
Choose the request that must continue
Start with a specific task, such as receiving a fictional maintenance request. Define what counts as receipt, who takes ownership and which information allows routing. Separate initial recording from resolution of the problem. Telling someone their request arrived is different from confirming that a team has performed the service.
The zone-loss recovery analysis helps map dependencies of the business function. Here, testing focuses on the handoff between channels and queues. Record available and suspended operations; a homepage failure does not describe every service.
Confirm the alternative channel’s dependencies
List the resources needed for service: authorized people, request identification, service information and access to a temporary record. Ask whether any depend on the same unavailable component. An instruction to call may be insufficient if the operator cannot record or track the request received through that route.
The technical-team continuity review complements preparation of owners and substitutes. Use individual identities and task-limited access. The procedure should explain who may consult records, correct routing and decide exceptions, without shared credentials or an improvised spreadsheet accessible to everyone in the organization, including people outside the approved service role.
Follow the request through a controlled queue
Prepare an authorized exercise with a simulated portal, an alternative intake route and synthetic requests. Agree on the simulated outage and stopping conditions. Do not interrupt an actual service to demonstrate the problem. Submit a normal request and another representing the same issue received through two different channels.
Observe these points before accepting the procedure:
- Receipt: assign an identifier and record when and through which channel the request arrived.
- Routing: connect the request to an owner and a status the next team can understand.
- Duplication: check how the team recognizes two entries concerning the same case without deleting necessary history.
- Follow-up: confirm how the test requester obtains an answer through the approved channel without exposing other people’s requests.
Also test the return to the primary channel
Define who authorizes the transition from the temporary queue to normal operations. Check which requests remain open, which were resolved and which need review. Connect identifiers used in both environments so the channel change does not create a second job or make a request disappear from the service record.
In the laboratory, restore the simulated portal and perform that transition. Compare records before and after, including the state of a requester who already received an answer. Record conflicts, external dependencies and manual decisions. The conclusion should describe reconciled requests and remaining limitations, with owners for resolving differences before closing the exercise.
Communicate the service and the next step
A useful message identifies the affected function, the known channel for continuing and how to follow an existing request. Avoid asking users to send documents or credentials through an improvised route. The portal login-journey analysis complements review of destinations and failure communication, keeping the instructions connected to the approved workflow.
Retain the sample, results and points requiring correction. Repeat validation when the supplier, authentication, service channel or queue integration changes. Acceptance covers the tested path; investigating a real incident and confirming current availability need their own decisions and evidence.
Frequently asked questions
Did the Columbia incident interrupt every service?
The notice distinguishes affected services from functions that remained operational. Do not use one channel’s outage to describe the entire administration or as a measurement of today’s conditions.
Does an interruption establish data theft?
Unavailability does not itself establish information access or copying. That conclusion depends on investigation, records and its own observation limits.
Does publishing an alternative phone number complete the plan?
The exercise should demonstrate how a request is received, assigned, tracked and reconciled. The published contact is only the start of that path.
Must the production portal be switched off?
The proposal uses simulated components and fictional data in an authorized environment. If impact is undefined, begin with a tabletop exercise and record that limitation.
Sources and references
Original Pentesty analysis, published October 23, 2026. The proposed scenarios are editorial guidance, with scope and validation to be defined for each environment.