ACME HQDemo work items and pages, made up for Emargen's demo. Not a real company.
Security / Security

Incident response

By Mira Holm · created 6 Mar 2023, 09:40 UTC · last updated 16 Sep 2026, 13:00 UTC · version 8

incident-responsepolicysecurity

Owner: Mira Holm · Reviewed every year; rehearsed twice a year.

What we do when something goes wrong in production or with customers' data. Operations runbooks are in the Engineering and Operations spaces; this page is the policy.

Is it an incident?

Yes, if any of these:

Not sure? Treat it as one. It is cheaper to close an incident than to open it late.

Severity

SeverityMeansExample
SEV1Many customers or a whole region affected, or a confirmed data breachSign-in down in the EU region
SEV2One customer or one feature badly affectedSMS codes delayed for hours (OPS-33)
SEV3Small or partly affected, with a way roundOne depot's single sign-on failing

Roles

Steps

  1. Open an OPS incident in Jira and say the severity.
  2. Stop the harm first; find the cause later.
  3. Status page within 15 minutes for SEV1, and for SEV2 as soon as customers would notice; then every 30 minutes until resolved.
  4. For a personal data breach: we tell affected customers within 48 hours of confirming it. Our DPA commits to 72.
  5. Close the incident when the harm has stopped and customers are told.
  6. Review within five working days, with Emargen's Incident review template.

Rehearsals

DateScenarioNotes
12 Mar 2025A laptop with production access is stolenRevoking AWS and GitHub access took 25 minutes
17 Sep 2025Backups can't be restored in one regionWho decides on a restore, and who tells customers, written down
18 Mar 2026Mail provider sends notices to the wrong customerCustomer lead role written down
16 Sep 2026Credential stuffing, after OPS-41Counted as the second rehearsal of 2026; added the attack runbook