Operational guardrails let you define rules about when and how changes should happen. When an event violates a guardrail, OpsTrails detects it and sends notifications to the channels you configure.
Guardrails are advisory — they detect violations and notify your team, but they do not block events from being recorded. Every event is always stored regardless of whether it violates a guardrail. This means your operational timeline remains complete while your team is alerted to potentially risky changes.
Guardrails evaluate against the following event types: deployment, rollback, incident, data-load, and deployment-status. See Core Concepts for more on event types, subjects, and sources.
A change freeze window defines a time period during which changes should not occur — for example, during a holiday weekend or a critical business event. If an event is recorded during an active freeze, the freeze is marked as violated and notifications are sent.
Deployment policies define ongoing rules for how deployments should happen — for example, no deployments on weekends, only during business hours, or no more than 5 per day. Policies evaluate continuously against incoming events.
Both freeze windows and deployment policies can send alerts through notification channels — email, webhooks, PagerDuty, Opsgenie, or Google Chat. Email notifications to organization owners and admins are always enabled by default.