Past incidents

1. Webhook delivery delays: resolved

   Sep 30

   A queue worker ran out of memory after a deploy. We rolled back and drained the backlog; no events were lost.

2. Elevated API latency

   Sep 18

   p95 latency rose to 1.2s for 18 minutes during a database failover.

3. Card payment errors

   Aug 20

   An upstream card network returned errors for 9 minutes. Retries succeeded.