SEV-2
Resolved
2 June 2026 · duration 18 min
Brief login delays for a subset of users
A configuration change to our session cache caused slow sign-ins for roughly 6% of users for 18 minutes. No data was exposed, and no money movement was affected, the read-only and approval paths were untouched.
Timeline (WAT)
09:14Automated alert: elevated login latency.
09:19On-call engineer identifies the cache config change.
09:28Change rolled back; latency returns to normal.
09:32Recovery confirmed across all regions.
Root cause & fix
A cache eviction setting was tuned too aggressively, increasing cache misses. We reverted it, added a latency guardrail to the deployment pipeline that blocks similar changes, and expanded pre-production load testing for session services.
SEV-3
Resolved
21 May 2026 · duration 42 min
Degraded account connectivity via a provider
One of our open-banking connectivity providers experienced an outage, so refreshing balances for some institutions was delayed. Cached data remained visible, and you could still use Qoyla; only live refresh was affected.
Timeline (WAT)
14:03Provider error rate climbs; we flag connectivity as degraded on status.
14:21We fail over affected institutions to a secondary route.
14:45Provider recovers; we monitor for stability.
Root cause & fix
The incident originated entirely with a third-party provider. We have since shortened our automatic failover threshold and added a clearer in-app banner so users know when a refresh delay is upstream, not on our side.
Maintenance
Completed
9 May 2026 · planned
Planned database maintenance
A scheduled upgrade to our primary database, announced five days in advance. Completed within the window with no user-facing downtime, thanks to a rolling, zero-downtime migration.
Outcome
Storage headroom and query performance improved. We mention planned maintenance here too, so this page is a complete record, not just a list of things that went wrong.