Our AWS-hosted infrastructure was unreachable for customers in Australia from roughly 8:00 to 9:00 AEST, then recovered without any changes on our side. We could still access the site ourselves, but our clients could not. Around the same time, outage reports for several major Australian services appeared to spike and then return to normal. Has anyone else seen a similar incident, or found an explanation for what caused it?
3 Answers
We eventually got confirmation from Imperva/Thales that one of their Melbourne proxies was offline and came back at about 8:44 AEST. It wasn't listed on the public status page, and we had to contact our account manager. They said a proper root-cause report would take a few days.
If you use Imperva, that may be the explanation. We heard that one of their Melbourne points of presence went offline, which would affect traffic passing through that location even if the origin infrastructure remained healthy.
Our monitoring showed a brief outage of roughly two minutes in the ap-northeast-1 region around the same general time, although that may have been a separate event. We didn't see a broad outage internally, so the regional proxy issue seems more consistent with customers being unable to connect while your own access still worked.
Yes, we do use Imperva. Their public status page didn't show anything at first, so I was trying to find independent confirmation.