Emergency Change Process: Approving an Urgent Change Safely
An emergency change is a change that cannot wait for the normal route. It needs a faster path, but not an absent one. The safeguards you keep are the ones that are cheap and effective.
An emergency change is an urgent change to restore service or to address a serious security issue. It cannot wait for the next scheduled review of changes. In ITIL practice it follows a shortened route, often with an emergency change advisory board that is a subset of the full board, and it is reviewed in full after the event.
Define what qualifies
The definition must be narrow. An emergency change responds to a service that is down or badly degraded, a security vulnerability being exploited or about to be, or a legal or regulatory duty with an imminent deadline. The need for it must come from a risk that grows materially with waiting. A late project, a customer demand and a manager's impatience are not emergencies.
The shortened route
- Raise it quickly, in writing, in the same record the incident uses if there is one. State the reason it cannot wait.
- Name the approvers in advance for out-of-hours, with deputies. Ask them for a decision, not an acknowledgement.
- Keep the minimum safeguards: a second person reviews the change, there is a way to undo it, and someone monitors the result.
- Implement with a person recording the steps and times as they are done.
- Verify against the symptom that caused the emergency.
Collisions with freezes
Emergency changes often meet change freezes. A freeze policy should name this route as its exception, so that the emergency does not need an argument. In Atlas, releasing to Production while a freeze covers it is refused unless a workspace owner or administrator overrides it with a recorded reason of up to 500 characters, which keeps the exception accountable.
The review afterwards
The cost of speed is paid back in review. Within a few days, examine each emergency change. Was it justified? Did it work? Did it cause a further problem? Could the need have been foreseen? Emergency changes that recur for the same reason point to a problem that should be recorded and fixed. See the problem management guide.
Track the proportion of changes that are emergencies. A rising share means that planning, testing or the normal route is failing.