Evidence kitevidence-folder-template/02-logging-monitoring/alert-routing.md
Logging & monitoring

Alert routing

# Alert routing

Evidence that security-relevant events route to a human (or documented runbook) within minutes. "We have alerts configured" is not enough — the questionnaire wants proof of routing and acknowledgement.

Evidence to keep here

  1. Alert rules — list of security-relevant alerts with their triggers, severities, and destinations.
  2. Routing config — PagerDuty / Opsgenie / native cloud alert integration that shows messages reaching a person.
  3. Sample fires — 1-3 anonymized examples of alerts that actually fired, were acknowledged, and resolved. Demonstrates the chain works.

Alerts that customer security reviews expect

At minimum:

Trigger Severity Destination
Root / break-glass account login Critical Page on-call
New IAM admin role assignment High Page on-call
New service-account key created in prod High Page on-call
Audit logging configuration changed Critical Page on-call
Multiple failed sign-ins on admin account High Page on-call
Public S3 bucket / GCS bucket created High Page on-call
Secret-scanning alert (committed credential) High Page security
Container image with critical CVE deployed High Page platform

Add more for your stack. Each item should have:

  • A configured alert rule (link / file path)
  • A clear destination (specific channel / person / rotation)
  • A documented runbook for the on-call to follow

How to gather the evidence

For each alert, screenshot or export:

  • The rule definition in {{Datadog / CloudWatch / Sentinel / GCP Alerting}}
  • The destination configuration in {{PagerDuty / Opsgenie}}
  • A sample firing (anonymize sensitive data)

Save as alert-<short-name>-<date>.{png,json}.

Sample answer for the questionnaire

Security-relevant events from cloud audit logs, IdP audit logs, and code-repository security alerts route to our on-call rotation via {{PagerDuty / Opsgenie}}. Alert rules and runbooks are listed in alert-routing.md. Triage during on-call follows the playbook in ../05-backups-recovery/incident-runbook.md.

Refresh

  • Alert config: annually unless you add new alerts.
  • Sample fires: include any meaningful examples that show the chain works (don't manufacture them, but if real alerts fired, file an anonymized sample).
Book review