minormonitoringStarted September 1, 2026 at 2:18 PM UTC
Search shard rebalance increased lookup time
Index Sweep nodes in Sydney crossed the normal latency band during a shard rebalance. Queries completed, but local mirrors took longer to settle.
Incident Archive
Timeline of what bent, slowed, and recovered.
Archive view for public incident history as of September 1, 2026. Exact UTC timestamps stay visible so operational drift is traceable, not hand-waved.
Current summary
Overall
Partial degradation
Generated
September 1, 2026 at 6:42 PM UTC
Recent pulse
5 systems normal, 1 monitoring, 1 degraded.
Event ledger
September 1, 2026 at 6:42 PM UTC
Status pulse
5 systems normal, 1 monitoring, 1 degraded.
September 1, 2026 at 4:05 PM UTC
Search rebalanced
Query fanout moved away from the overloaded Sydney shard pair.
August 31, 2026 at 8:47 AM UTC
Calendar resolved
Recurring expansion restored after cache keyspace rebuild.
August 28, 2026 at 11:03 PM UTC
Dispatch recovered
Retry latency returned to the standard band.
majorresolvedStarted August 31, 2026 at 6:02 AM UTCResolved August 31, 2026 at 8:47 AM UTC
Recurring event expansion stalled for one region
Calendar Mesh workers in Portland stopped expanding recurring events after a stale cache key survived rollout. The queue drained after the keyspace was rebuilt.
minorresolvedStarted August 28, 2026 at 10:11 PM UTCResolved August 28, 2026 at 11:03 PM UTC
Outbound retry window widened under provider backpressure
Outbound Dispatch held messages longer than expected while upstream backpressure cleared. No data loss occurred and retries completed automatically.