Changelog

What we shipped

Everything customer-facing, newest first. Nothing here needs an upgrade on your side, and nothing here changed a price.

Subscribe to this page as a feed

  1. Subscribers can now follow a single service

    Added

    A subscriber picks the services they depend on instead of taking every notice the page sends. Existing subscribers keep everything until they choose otherwise, so nobody stops hearing from you by accident.

  2. Incident updates keep the time they were written

    Changed

    Editing an update no longer moves it to the top of the timeline. The original time stands and the edit is noted underneath, so a review afterwards reads in the order things actually happened.

  3. Certificate renewal failures now page someone

    Fixed

    A renewal on a custom domain could fail quietly and keep failing. It now raises an alert on the first failure and emails the account owner on the third.

  4. Templates for the first update

    Added

    Save the wording your team already agreed on and open an incident from it. Placeholders for the service, the impact and the time of the next update are filled in for you.

  5. Uptime is counted one way everywhere

    Changed

    Partial degradation now counts as partial in the dashboard, the monthly report and the API, using the definition written on the reporting page. Historic figures were recalculated and may differ slightly from what you saw before.

  6. Long incident titles no longer overflow on a phone

    Fixed

    A title with a long unbroken token could push the public page sideways at narrow widths. It wraps now, and the subscribe control stays where it belongs.

Older entries live in the archive. Security notices are published separately, on the security page.