Work Culture & Behind the Scenes

How the UK Ministry of Justice Platform Team Handles Incidents: Declaration, Lead and Scribe Roles, and Postmortems

The UK Ministry of Justice's Modernisation Platform runbook shows how to handle incidents in a structured way, from definitions and P1–P3 priorities to clearly defined roles.

By Oemah Solution Editorial 4 min read

This article was prepared with AI assistance from the official sources listed below, and checked automatically before publication.

An empty, long control room with rows of panels, buttons, and indicators, symbolising system monitoring
Photo: Thurmond Dam Control Room (9272094428) by U.S. Army Corps of Engineers Savannah District · Public domain · Wikimedia Commons

Incidents on digital services are often handled spontaneously: someone reports in a group chat, many people chip in, and no one clearly leads or keeps records. As a result, decisions are scattered and lessons are lost. The Modernisation Platform team at the UK Ministry of Justice has set out its process for handling network, operational, and security incidents in a public runbook. This article summarises its contents as study material.

What counts as an incident and which roles are mandatory

According to the runbook, an incident is an unplanned event that meets one of the following conditions: it affects end users or direct users (developers and engineers), it degrades the quality of the service users see, or it increases risk to a production service. If an event does not meet that definition, the recommended step is usually to log an issue in a GitHub repository. If it is security-related, the issue is logged in a dedicated security repository.

Two roles must be filled for every incident:

  • Incident Lead: coordinates the response, makes sure all tasks are being carried out (rather than doing everything personally), makes the final decisions, decides whether additional roles such as a communications lead are needed, declares the incident resolved, and ensures the post-incident process is carried out.
  • Scribe: keeps a log of key events, discussion topics, decisions, actions, and investigation findings. This log is not a word-for-word transcript, but a summary such as who proposed what and who is following up.

The runbook acknowledges that one person can hold both roles in rare cases, but does not recommend it because it generally leads to poor record-keeping. Volunteers are requested from the team, and if there are none, someone is appointed.

How the process works

The flow in brief: confirm the incident, declare it, assign roles, manage it, fix it, then follow the post-incident procedure.

  • Confirmation. If the incident is a security breach, the team must also contact the Service Desk to have it logged.
  • Declaration. From Slack, the /pd trigger command opens the Declare Incident form in PagerDuty. It covers a short title, incident type, affected service (network, operational, or security), an optional description, urgency, and priority. A dedicated Slack channel for the incident is then created.
  • Priority. Only three levels are used. P1: the entire platform is down. P2: part of the platform is down and some user applications are affected. P3: part of the platform is down, but user applications are still available.
  • Communication. For P1 or P2, the Head of Platforms and Architecture, the Head of Hosting, and the Product Manager/Delivery Manager must be informed, in addition to service users. External status updates are published through the PagerDuty status page, and the runbook stresses that this must not be replaced by manual notifications.
  • Role handover. In long incidents, the role holder is responsible for making sure there is a replacement, and the handover is announced in the incident channel.
  • Fixing. Not every incident needs the whole team, even though everyone wants to get involved.

Pitfalls to avoid

The runbook reveals several pitfalls it deliberately guards against:

  • Choosing the wrong affected service in the form. If a service outside the three options is selected, notifications to the high-priority channel and the status integration do not run automatically.
  • Announcing only via chat without updating the external status page, even though that page feeds other services that depend on the platform.
  • An incident lead who dives in and does everything personally, so coordination is neglected.
  • Combining the Lead and Scribe roles, resulting in poor records.
  • Writing the log as a long transcript rather than a summary of decisions and actions.

Post-incident and how to apply it

After an incident is closed, the runbook asks for three things: creating a new incident report, scheduling a blameless postmortem meeting, and publishing a runbook to fix similar problems. The document itself is reviewed periodically; the last review was on 10 November 2026 and the next is scheduled for 10 February 2027.

Teams and clients can adopt this pattern as a lightweight template:

  • Write an incident definition agreed by everyone, so it is clear when an event needs to be declared.
  • Set a simple priority scale with concrete impact criteria.
  • Appoint an Incident Lead and Scribe from the start, along with a list of parties who must be informed at high priority.
  • Make sure the reporting channel for security incidents to the competent authorities is included in the procedure.
  • Schedule a blameless postmortem and turn its outcomes into a reusable runbook.
  • Review the process document periodically so it stays relevant.

The source is just one runbook, but the principle applies broadly: incidents that have a definition, roles, and records are easier to control and learn from.

#incident-management#runbook#postmortem#devops#sre

Sources & references

This article is a summary and interpretation by the Oemah Solution team of the sources below. Refer to the original sources for full details and context.

  1. Incident Process Ministry of Justice (UK) – Modernisation Platform · user-guide.modernisation-platform.service.justice.gov.uk