Skip to content

D4 · Incident response

This plan defines how a security incident at Bolívar Connect is detected, contained, analyzed, fixed and communicated, and who does what.

Any event that compromises or may compromise the confidentiality, integrity or availability of the platform or its data. For example:

  • use of an API key by someone who should not have it, or an exposed credential;
  • a tenant accessing another tenant’s data;
  • duplicated, altered or unauthorized payment instructions;
  • exposure of personal data;
  • an outage of the service or of a regulated financial institution;
  • anomalous activity in the infrastructure.
Level Criterion Initial response
SEV1 — Critical Personal data exposed, unauthorized instructions, cross-tenant access or service down Immediate, at any hour
SEV2 — High Compromised credential without confirmed use, severe degradation or a third-party failure Within 4 hours
SEV3 — Medium or low Suspicious event without confirmed impact On the next business day
Role Who Function
Incident coordinator Security Officer Declares the incident, sets the severity, decides and coordinates
Technical team On-call technical personnel Contains, analyzes and fixes
Communications Management Notifies clients, third parties and authorities when appropriate

Anyone on the team who detects a possible incident reports it immediately to the Coordinator.

Sources: anomalous authentication and rate-limiting rejections, third-party circuit breaker errors, reconciliation differences, notices from clients or third parties, and AWS alerts. An incident record is opened with the time of detection, the source and the initial severity.

Situation Action
Compromised API key Immediate revocation and issuance of a new one to the client
Exposed webhook secret Secret rotation
Exposed third-party or infrastructure credential Rotation in AWS Secrets Manager and service restart
Malicious traffic IP restriction in the security group
Suspicious beneficiary Beneficiary blocked
Failing third party Circuit breaker, which takes it out of service
Server compromise Instance isolated and recreated from infrastructure as code

Evidence is preserved before any destructive change: logs, snapshots and the audit log.

What happened is reconstructed using each request’s request_id, the audit log, the status history of each payment instruction and the AWS logs. The root cause, the scope (affected tenants, instructions and data) and the time window are determined.

The root cause is fixed with a regression test and the service is validated. If payment instructions were affected, they are reconciled against the regulated financial institution’s records before the incident is closed.

To whom When What
Affected clients Within 24 hours of confirming a SEV1 or SEV2 that affects them Scope, impact, measures taken and next update
Regulated financial institution When the incident affects instructions sent to it Affected instructions and required action
Data protection authority Within 72 hours when there is a personal data breach subject to the GDPR or applicable regulation What the regulation requires

Within 10 business days after closing a SEV1 or SEV2, a report is written with the timeline, root cause, impact, corrective actions and their owners. The actions are followed up until closed.

The plan is tested at least once a year with a tabletop exercise on a realistic scenario, for example a leaked API key or a third-party outage. The result and improvements are recorded.

The escalation contact list (team, Management, third parties and clients) is kept up to date outside this document and reviewed at every plan test.