System status · Refreshed when something changes

All channels open.

Last checked · 2026-05-22 14:02 UTC

Six services. Ninety days of uptime to look at. If something is broken, it'll say so here in a complete sentence.

Keep an eye on the barn.

RSS →History →

Services

What's in the barn, and how it's doing.

Six services. Ninety days behind each one. The number on the right is uptime, rolling. The dots are the last ninety days, oldest on the left.

  • Dashboard & API

    Operational.

    99.98%

     

  • Email channel

    Operational.

    99.96%

     

  • Discord bridge

    Operational.

    99.91%

     

  • Slack bridge

    Operational.

    99.93%

     

  • Webhook delivery

    Operational.

    99.89%

     

  • Marketing site

    Operational.

    100.00%

     

nothing happeneddegradationoutage

Incident journal

The honest log.

2026-04-18degradation

Discord bridge held its breath for nineteen minutes.

A rate-limit window from Discord we hadn't seen before; our retry budget burned in seven minutes and inbound DMs queued behind a slow worker. Theo restored the worker pool at 14:07 UTC; backlog drained by 14:11. We've added a separate retry queue and lowered the budget alarm threshold from 80% to 50%.

2026-03-02outage

Eight minutes where nobody could log in.

A config rollout removed an environment variable the dashboard reads at boot. The deploy passed CI because the test suite mocks the variable. We rolled back at 09:48 UTC; sessions were unaffected (only the login endpoint was down). Maren added a startup assertion and a smoke-test on the real env.

2026-01-30maintenance

Database migration, 03:00–03:42 UTC.

Planned move to a new Postgres major version; we drained writes for 42 minutes. No customer reported impact. Notes for next time: 04:00 UTC suits our European customers better than 03:00.

If you're seeing something we aren't — send us a note.

↳ All timestamps in UTC.