Incidents and maintenance

The incident composer, states and impact, templates, update progression and resolving, scheduled maintenance with automatic start and completion, postmortems, and subscriber notifications.

Incidents are posted from the Open incident button in the workspace and render as a timeline on the public page. Each incident has a title, an impact, a current state, affected components, and one or more updates over time.

The composer#

One panel handles everything, and it works one-handed on a phone:

1. Pick Incident or Maintenance.
2. Tap a template if you have one; it fills in the title, message, impact, and components.
3. Give it a name.
4. Choose the status and the impact (none, minor, major, critical).
5. The message is prefilled with a sensible sentence for the status; edit it or leave it.
6. Tick the affected components. Each one gets a status that follows the impact you chose (minor sets Degraded, major sets Partial Outage, critical sets Outage, maintenance sets Maintenance), which you can change per component.
7. If visitors can subscribe to your page, choose whether to notify subscribers.

Press Cmd/Ctrl+Enter or the button. A new incident with a template is a few seconds of work.

Incident states#

An incident moves through these states. The labels are renamable per page; the colors and icons are fixed so the timeline reads consistently:

Posting updates#

From an open incident, tap Post update. The composer opens with the next state already selected (Investigating leads to Identified, then Monitoring, then Resolved) and a message prefilled for it, so a routine update is one or two taps. You can change the status of the affected components in the same step, or add more components. You can also edit an update's text afterwards, or delete one you posted by mistake; the opening update stays.

Resolving#

When you move an incident to Resolved, the composer shows every component the incident touched with a "restore to Online" tick. Components that another open incident is still affecting are left alone and marked as such, so resolving one incident never hides a second problem. Resolving ends the downtime those components contributed to uptime.

Templates#

Save reusable templates (name, title, message, state, impact, and the components to affect, with the status to set them to) and they appear as chips at the top of the composer, most used first. You can also save any incident you are about to post as a template.

Scheduled maintenance#

Maintenance is an incident of its own kind with a start and end window. When you schedule it for later:

• It appears under Scheduled Maintenance on the public page straight away, with the window shown in UTC.
• Nothing changes on your components until the start time.
• With auto-start on, Hep.gg sets the affected components to Maintenance and posts an "in progress" update at the start time; you can also press Start now.
• With auto-complete on, it posts a completion update and restores the components at the end time; you can also press Complete now.
• With remind subscribers on, subscribers get a reminder one hour before it starts.

You can edit the window or cancel a maintenance before it starts. Maintenance time does not count as downtime in the uptime bars.

Postmortems#

After an incident is resolved you can write a postmortem and publish it. It renders under the incident in the page history, and you can notify subscribers when you publish.

Notifying subscribers#

Visitors can subscribe to your page by email (you turn this on under Settings, Notifications), and you can add webhook and Discord channels of your own under the Subscribers view. Each incident, update, maintenance event and published postmortem can notify them; monitor-opened incidents notify automatically unless you turn that off. Every page also publishes Atom and RSS feeds of its updates. Details in Public endpoints.