1 service is reporting reduced performance. Everything else is running normally. Figures below cover the trailing 90 days.
| Service | Status | 90-day uptime |
|---|---|---|
| API | Operational | 99.99% |
| Console | Operational | 99.98% |
| Provisioning - US East | Operational | 99.96% |
| Provisioning - US West | Operational | 99.95% |
| Provisioning - EU West | Operational | 99.97% |
| Provisioning - AP South | Degraded | 99.71% |
| Block storage | Operational | 99.99% |
| Object storage | Operational | 99.99% |
| Network fabric | Operational | 99.98% |
| Billing | Operational | 100% |
Queue times for new instances in ap-south-1 are averaging four minutes against a normal median of 40 seconds. Running instances are unaffected. We have identified a scheduler placement issue and are rolling out a fix region by region.
A session store migration caused intermittent login failures for roughly 25 minutes. The API was unaffected throughout. Sessions created during the window were invalidated; no credential data was exposed.
A failed optical transceiver in the eu-west-1 spine reduced non-blocking capacity for 90 minutes. Multi-node jobs saw reduced all-reduce throughput; single-node work was unaffected. Affected customers were credited automatically.
Hosts drained progressively. Running instances are not interrupted.