fra-01 — brief BGP route flapFrankfurt PoP saw a transient BGP route flap with our DE-CIX transit provider. Failed requests at <0.01%; p50 latency briefly elevated to ~38 ms before traffic was rebalanced through ams-01. Root cause was a misconfigured route reflector on the transit side, confirmed by DE-CIX.
per-01 after disk replacementHot-replaced an SSD in the Perth PoP. The replacement node took longer than expected to rejoin the KV write quorum (14 min vs the usual ~3). Reads were unaffected (served from other PoPs). KV writes returned 503 at Perth and were correctly retried by client SDKs.
A dashboard deploy queued behind a long-running migration. Sign-in worked but was slow. Functions, KV, and Objects were unaffected throughout. Migration runtime is now bounded.
One email when an incident opens, one when it resolves. No marketing, ever.