Status

Every minute we were down, and why.

Uptime is measured from outside, from three regions, on the enqueue endpoint. A minute counts as down if two of the three probes fail.

Enqueue API

operational

Workers and leases

operational

Dashboard

operational

Schedules

operational

Webhooks out

degraded

Last six months

April 2026

100%

0 min down

May 2026

99.99%

4 min down

June 2026

100%

0 min down

July 2026

99.94%

26 min down

August 2026

100%

0 min down

September 2026

99.99%

3 min down

Every incident since we opened

14 July 2026

26 min

Enqueue returned 503 in Frankfurt

A migration held a lock on the jobs table for longer than the deploy timeout. Jobs queued in the SDK buffer and drained after. Nothing was lost. We now run migrations behind a lock timeout of two seconds.

3 May 2026

4 min

Dashboard would not load

A bad build reached production because the smoke test only checked the API. Rolled back in four minutes. The smoke test now opens the dashboard.

11 September 2026

3 min

Webhooks out were delayed

One region ran out of outbound connections during a spike. Queued and delivered inside three minutes. The pool is now sized to three times peak.