Upstream Cin7 Core API rate-limit changes caused our sync to back off and queue. We adjusted our batching strategy and rolled out a hotfix. No data loss; all queued syncs caught up by 10:02.
Planned RDS upgrade triggered a 41-minute read-only window for the web app. Forecast and reorder calculations continued in the background; no customer action required after recovery.
17-minute lag on webhook processing for ~6% of workspaces. Caused by a noisy-neighbour deploy. Backfill ran automatically.