'Service status' pattern

[paulmsmith]

What

A pattern for helping users understand the 'status' of a service. - For example the service's availability during a planned or unplanned outage.

Why

All services encounter times when part of (or the entire service) is down, either for maintenance or because something has gone wrong.

Services should be prepared for when this happens and I believe there are common user needs within a user's experience of this scenario.

Users need to know things like

1. If the service knows about the issue or not.
2. How to report issues if and when they encounter them.
3. How they might be impacted.
4. Estimate time to a resolution/fix.

Examples

• GOVUK Wifi Status
• GOVUK Notify Status
• Planning Data service status

Related links:

• Monitoring the status of your service - Service Manual
• When everything's gone wrong - 'Becoming a teacher' design history.

[paulmsmith]

My team encountered an unplanned outage due to a problem with DNS configuration which meant the service could be intermittently unavailable for some users depending on how and where they were accessing the internet.

I've suggested this as a pattern because I think it applies to many potential services and possibly encompasses and but goes further than some of the existing content/page patterns such as Service unavailable pattern within the design system currently.

[edwardhorsford]

We recently blogged about our 'Everything's gone wrong' page - that we updated quickly during an incident to include a timestamped update message with the current status.

Our outage lasted 4.5 hours and we ultimately added 2 update messages.

One thing we cover in the thing is how our page is somewhat separate from our other infrastructure.

[paulmsmith]

I've noticed already in our initial version we've defaulted to quite technical language such as 'incident' and 'outage'. We did it fast but I think if people have research or good content patterns that might help others.

[paulmsmith]

The team have since added an 'Incident response' in our documentation about the issue for those that might want more context and what led to us developing a service-status aspect to our service.