OVPN Transparency report January 2016
We didn't publish any transparency reports for November and December since we moved to a new datacenter and also changed our internal system to keep track of the traffic volume. Going forward, we will continue publishing monthly transparency reports detailing the traffic volume and load.
Since we've increased the number of servers, we're going to start reporting statistics per datacenter instead of per server - since that gives a better and more accurate overview of our system load.
OVPN had one service interruption in January that affected servers in Stockholm. The interruption lasted for roughly ten minutes and was caused by a DDoS aimed towards the hosting provider.
Our servers were online at least 99,5% of the time in January. OVPN also had a new traffic record, since we encrypted and anonymized 806 TB.
Check below for more statistics and raw data.
VPN servers in Malmö, Sweden
Traffic peaks (Mbit/s)
The first picture has been adjusted according to the actual traffic volume, and the second after our capacity. OVPN has enough capacity to deliver 10 Gbit/s in Malmö, but the average traffic volume in January was 483 Mbit/s.
The highest traffic peak was 1280 Mbit/s, which is 12,8% of the max capacity. The 95th percentile was 838 Mbit/s, which means that 95 percent of the time, the servers in Malmö had a load of 8,4%.
Malmö had an uptime of 100% in January, even though there's a blank part in the graph.
VPN servers in Stockholm, Sweden
The first picture has been adjusted according to the actual traffic volume, and the second after our capacity. OVPN has enough capacity to delivery 10 Gbit/s in Stockholm, but the average traffic volume in January was 1930 Mbit/s.
The highest traffic peak was 3750 Mbit/s, which is 37,5% of the max capacity. We had a short outgoing burst of 5770 Mbit/s on January 4th since OVPN was used to participate in a DDoS. This attack was quickly mitigated.
The 95th percentile was 2860 Mbit/s, which means that 95 percent of the time, the servers in Stockholm had a load of 28,6%.
Stockholm had an uptime of 99,99% in January.
WEB01 had an uptime of 99,96% in January, which means we had a down time of 16 minutes. The average response time was 280 ms.
SQL01 had an uptime of 99,98% in January, which means we had a down time of 7 minutes.