Opened on Sunday 18th August 2019, last updated
Resolved — It appears the issue has been resolved on August 22th at 8pm (UTC+2) and the situation has been stable since then. Still without an explanation from OVH though, we finally got a first response on the 27th but nothing more than basic questions to deflect the report.
Investigating — Since August 4th at 6:10am (UTC+2) we've noticed a big increase in network errors between our Roubaix (OVH) location and many destination IPs (mostly Cloudflare, OVH and more). This shows up mostly as random "SSL connect error" and "Failure when receiving data from peer". As it only impacts one location this should not generate false positive alerts as the double check will have to fail too, but it may show up on your status page (and on your dashboard if you look at the right time) as failed requests. Our hosting provider for this location (OVH) has been contacted on August 7th but no single response so far... We'll keep you updated. In the meantime if you want to avoid this you should remove the Roubaix location from your impacted check (and put it back once resolved).