We were alerted to small outages occurring on our infrastructure in US West at 2.30am (GMT+12). An initial investigation determined that the infrastructure was online and that this micro outage was no longer in place. We continued to see these alerts and micro-outages through the period until 7.21am (GMT+12) . We have spun up more infrastructure in the region to accommodate load across our application servers which was causing a memory limit to be hit for small periods of time.