How to manage technical incidents

GDS incident management focuses on restoring normal operations quickly with minimal impact on users.

Technical incidents might also be cyber security or data loss incidents. You must report all suspected or actual cyber security incidents to the CO:D Cyber Security team and to the GDS Information Security team using the Incident response process. You must report all actual or suspected data breach incidents to the GDS Information Management team. These requirements should be included in your service manual/guides/processes.

Define incident priority

Define technical incident priority levels for your service’s applications. For example potential incidents include:

system access problems
wider technical failures with possible reputational impact to GDS
denial of service (DoS)
data breach or leak
defacement
unauthorised use of systems
suspicious activity, such as traffic from an unknown source

Assign a priority level to incidents based on their complexity, urgency and resolution time. Incident severity also determines response times and support level.

Example incident priority table

Classification	Type	Example	Response time	Update frequency
P1	Critical	Complete outage, or ongoing unauthorised access	20 minutes (office and out of hours)	1 hour
P2	Major	Substantial degradation of service	60 minutes (office and out of hours)	2 hours
P3	Significant	Users experiencing intermittent or degraded service due to platform issue	2 hours (office hours only)	Once after 2 business days
P4	Minor	Component failure that does not immediately impact a service, or an unsuccessful DoS attempt	1 business day (office hours only)	Once after 5 business days

Develop an incident workflow

Your team must understand what to do during an incident. Develop and document your incident workflow to reflect your service needs and team size.

Example workflow

Follow a prepared workflow to manage an incident to minimise its impact on your team and service users. Make sure that every step of the way is documented in writing using the incident report template [^1].

Establish an incident lead.
Inform your team.
Prioritise the incident.
Form an incident response team.
Investigate.
Contain.
Eradicate.
Recover.
Communicate to a wider audience.
Resolve the incident.

1. Establish an incident lead

Establish who your incident lead is. Find out who noticed the problem and if anyone else is investigating and fixing it. If that person is you, assume the role of incident lead.

2. Inform your team

Inform your team using your chosen tool, like Slack. If the incident involves a data or security breach, you must also notify:

The CO:D Cyber Security team. You can also use the #cyber-security-help Slack channel to contact CO:D Cyber within normal working hours.
The GDS Information Security team.
The GDS Information Management team.

3. Prioritise the incident

Prioritise the incident and start tracking actions, updates and communications. Teams like GOV.UK PaaS and Notify do this by creating a new incident report - copied from the incident report template - and use it to track updates and progress.

4. Form an incident response team

Form a team with both an incident lead and a communications lead. The communications lead will make sure relevant parties are updated according to the incident priority table.

5. Investigate

Make sure you keep your incident report up to date. If the incident involves a data breach follow your team’s GDPR documentation.

If the incident is a data or security breach you should follow steps 6, 7 and 8. If the incident is not cyber security-related, skip to step 9.

6. Contain

You should determine the right containment procedures. In some cases, you may require a forensic clone.

6.1 Short-term containment

You should start short-term containment measures as soon as you detect an incident. This could help minimise impact and maintain availability. Make sure that all affected systems are isolated from the non-affected systems.

6.2 Long-term containment

You’ll need to make sure long-term containment is in place.

You should take the system offline if possible. Once the system is offline, you can proceed to step 7.

If the system has to remain in production, remove all malware and other artifacts from the affected systems, and harden the affected systems from further attacks. You should reimage the affected systems, or restore from the last known good backup.

6.3 Forensic clone

As well as gathering evidence to help resolve an incident, you should collect evidence to support any potential follow-on disciplinary or legal proceedings.

To maintain the forensic integrity of the environment you should:

document all commands used during the investigation and keep the documentation up to date - include how the evidence has been preserved
store any forensic images taken during the investigation in a secure location, to prevent accidental damage or tampering

7. Eradicate

Eradication may be necessary to remove components of the incident that remain on your systems, such as traces of malware. To help with eradication you should:

identify all affected hosts
remove all malware and other artifacts left behind by the attackers
reimage and patch the affected system
check backups, code, images and the affected systems are protected against further attacks

8. Recover

Recovery is necessary to reduce the impact on user confidence and to reduce the likelihood of further successful attacks.

You should:

confirm the affected systems are patched and hardened against the recent attack, and possible future attacks
decide what day and time to restore the affected systems back into production (if they were taken offline)
check the systems you’re restoring to production are not compromised in the same way as the original incident
consider how long to monitor the restored systems for, and what to look out for

9. Communicate to a wider audience

If the incident is serious (P1 or P2) you’ll need to contact a wider GDS audience and potentially your service users.

Your communications lead must manage:

external and internal communications
incident escalations

External and internal communications

Make sure internal and external parties, like Information Security or your service users are fully informed at every stage of your incident management process.

For example, teams including GOV.UK Platform as a Service (PaaS), GOV.UK Notify and GOV.UK Pay use the StatusPage service to trigger notifications to subscribed users.

Post regular updates to the status of an incident in the #incident Slack channel. This helps people across GDS without having to find and follow multiple notification mechanisms for the different programmes.

Incident escalations

Notify internal escalation contacts of all high priority incidents (P1/P2). Contact the GDS Information Security team if you need help defining the escalation route for your service.

Report cyber security incidents

The incident lead, guided by the Information Security team, must inform the National Cyber Security Centre (NCSC) of any category 1, 2 or 3 incidents. The NCSC defines security incidents in its categorisation system prioritisation framework.

Depending on the incident, the NCSC may be able to provide technical support.

10. Resolve the incident

Hold an incident and lesson learned review following a blameless post mortem culture so your service can improve. Add a row to the central GDS incidents summary spreadsheet linking to your incident report document.

Example incident management

GOV.UK PaaS incident management process