An incident channel that is created, not chosen

Deciding where to coordinate during an incident is a decision made badly under pressure, and automating the creation removes it.

/incident "checkout returning 500s"

  → creates #inc-2022-06-14-checkout
  → invites the on-call, the service owner, the incident lead
  → pins a message with the runbook and the dashboard
  → starts a timeline document
  → posts a link in #engineering

and at the end: /resolve, which archives the channel and
attaches the timeline to the postmortem.

The timeline being captured automatically is the part that pays for itself, because reconstructing one afterwards from three channels and two direct message threads is most of the work of writing a postmortem. Archiving rather than deleting keeps the record searchable, which is how a similar incident eighteen months later gets resolved in ten minutes.