Statup
Language, English
GitHub
Use / Incidents

Use

Incidents

Declare what is broken, keep people posted with updates, resolve it, and write the report.

An incident says what is broken, which services it touches and how badly, and what the team is doing about it. Editors and administrators publish one from New event, on the dashboard or on the Events page.

Declare an incident

In the form, with Type set to Incident:

FieldDescription
ImpactRequired: Minor or Major.
Services affectedOptional. Services are listed under their folders, each folder with a box that ticks its services; past eight services, a search narrows the list, and the box then ticks only the services it shows. A line under the list says what will change, such as “Email: outage.”
TitleRequired, up to 200 characters. Until you type one, it writes itself, such as “Outage on Email”.
DescriptionOptional: what is affected, what is known.

Under More options:

  • Progress: Investigating by default, or Fixing or Monitoring.
  • Real start, if the incident began earlier than now. It cannot be in the future.
  • Next update: see below.
  • Save as a template: see Templates.

Publish the incident shows it at once, and sends it to the alert channels that receive incidents.

The description, the updates and the report take simple formatting: **bold**, *italic*, lists, [links](https://example.com), code and tables. No images.

From a detected outage

When monitoring finds a service down, Declare an incident, on the banner or the Services page, opens the form with the impact Major, that service, and the start of the outage as the real start.

What it does to the services

While the incident is Investigating or Fixing, the services it names show Outage for a Major impact, Degraded for a Minor one. From Monitoring on, they go back to their own state. See Services.

Progress

ProgressMeansCan move to
InvestigatingThe team is looking into it.Fixing, Monitoring, Resolved, Cancelled
FixingThe fix is under way.Monitoring, Resolved, Cancelled
MonitoringThe service is restored and watched. The timeline says “Service restored”.Fixing, Resolved
ResolvedOver. The incident shows how long it lasted.
CancelledClosed without counting in the services’ history.

Back to “Investigating”, or to whichever state came before, undoes the last move, one step, on the incident’s page.

Post an update

Post an update, on the incident’s page or in its side panel, takes a message, a new progress, or both, and the next update. Publish adds it to the timeline and sends it to the alert channels.

Each update shows its time. The name of its author is shown to signed-in members only.

Next update

Announcing when the next update will come tells readers when to look again.

  • A delay: 30, 30 min, 1 h, 1h30, up to 4 hours.
  • A time of day: 14:30, 14h30, 9h. It means the next time that hour comes, today or tomorrow, in the instance’s time zone.

The banner then says “update due at 14:30”, and “update due 5 min ago” once it is late. The incident’s page says “last update 9 min ago, next due at 14:30”.

Every update sets the next one again: left empty, it clears it. A delay is offered again in the following update; a time of day is not. Resolving or cancelling the incident drops it.

Reminder

When an update is more than 5 minutes late, the member who announced it gets a notification, “Your update is overdue”, and an email when a mail server is set. Once for each time announced.

Resolve

Choose Resolved in Post an update, with a message or without. The services go back to their own state, and the incident shows “lasted 2h”. A message posted within two minutes of the resolution shows with it, as one moment of the timeline.

Write the report

Once an incident is Resolved, Write the report: what happened, why, and what changes so it does not happen again. Publish the report shows it at the head of the incident’s page and in its side panel, marks the incident “Resolved · report” in the lists, adds it to the feed, and sends it once to the alert channels.

Edit, in the Report card, corrects it and sends nothing again; Remove takes it away. A cancelled incident has no report. If an administrator reopens the incident, its report is hidden until it is resolved again.

Edit or delete

Edit changes the title, the description, the impact and the services. The type, the real start and the next update stay as they are. An edit sends no alert.

Delete the event, at the foot of the edit page, removes it everywhere, its updates and report included, and sends no alert. One update can be deleted on its own: by its author while the incident is open, or by an administrator.

Once an incident is Resolved or Cancelled, only an administrator can still edit it, post on it, change its progress or delete it. Editors can still write its report.

Templates

A template keeps the title, the description, the type, the impact and the category of an event, without its services or times.

  • Tick Save as a template when publishing a new event.
  • Start from a template, at the top of the form, lists the 30 most used. Typing a title of two letters or more also suggests saved templates.
  • Applying one fills the title, the description, the type, the impact and the category. Services and times stay to be chosen.
  • A template is deleted from the Start from a template list.

Who can do what

ReaderEditorAdministrator
Read the incidentsYesYesYes
Declare, update, edit or delete an open incidentYesYes
Write, correct or remove the reportYesYes
Change a resolved or cancelled incidentYes